Calhoun: The NPS Institutional Archive
Theses and Dissertations Thesis Collection
1992-03
Multimedia electronic mail: standards and
performance simulation
Choi, Nag Jung
Monterey, California. Naval Postgraduate School
http://hdl.handle.net/10945/23975
dtt
NAVAL POSTGRADUATE SCHOOLMonterey, California
THESISMultimedia Electronic Mail: Standards
and
Performance Simulation
by
Choi, Nag Jung
March, 1992
Thesis Advisor: Myung W. Suh
Approved for public release; distribution is unlimited.
1257779
SECURITY CLASSIFICATION OF THIS PAGE
REPORT DOCUMENTATION PAGE
1a REPORT SECURITY CLASSIFICATION 1b RESTRICTIVE MARKINGS
2a SECURITY CLASSIFICATION AUTHORITY
2b DECLASSIFICATION/DOWNGRADING SCHEDULE
3 DISTRIBUTION/AVAILABILITY OF REPORT
Approved for public release; distribution is unlimited.
4 PERFORMING ORGANIZATION REPORT NUMBER(S) 5 MONITORING ORGANIZATION REPORT NUMBER(S)
6a NAME OF PERFORMING ORGANIZATION
Naval Postgraduate School
6b OFFICE SYMBOL(If applicable)
32
7a NAME OF MONITORING ORGANIZATION
Naval Postgraduate School
6c ADDRESS (City, State, and ZIP Code)
Monterey, CA 93943-5000
7b ADDRESS (City, State, and ZIP Code)
Monterey, CA 93943-5000
8a NAME OF FUNDING/SPONSORING
ORGANIZATION8b OFFICE SYMBOL
(If applicable)
9 PROCUREMENT INSTRUMENT IDENTIFICATION NUMBER
8c ADDRESS (City, State, and ZIP Code) 10 SOURCE OF FUNDING NUMBERS
program Element No Project N*. Work Unit Aae>iion
Number
1 1 . TITLE (Include Security Classification)
MULTIMEDIA ELECTRONIC MAIL: STANDARDS AND PERFORMANCE SIMULATION
12 PERSONAL AUTHOR(S) Choi, Nag Jung
13a TYPE OF REPORT
Master's Thesis
13b TIME COVERED
From To
14 DATE OF REPORT (year, month, day)
March 1992
15 PAGE COUNT
94
16 SUPPLEMENTARY NOTATION
The views expressed in this thesis are those of the author and do not reflect the official policy or position of the Department of Defense or the U.S
Government.
17 COSATI CODES
FIELD GROUP SUBGROUP
1 8 SUBJECT TERMS (continue on reverse if necessary and identify by block number)
Multimedia; Electronic Mail; Standards; Standardization; Simulation. LAN; ODA; CGM;JPEG; MPEG
19 ABSTRACT (continue on reverse if necessary and identify by block number)
This thesis surveys the current multimedia electronic mail (e-mail) related standards. The increasing demands for interoperability pushed the
international standardization organizations to develop standards for e-mail. The multimedia e-mail related standards include X.400 and ODA.ODA supports formatted text, raster graphics, and geometric graphics. The future standard will be able to support high bandwidth uses such as
high-resolution color still-image, full-motion video, voice, and audio. The future components of multimedia e-mail will include TIFT' for i aster
graphics, CGM for geometric graphics, JPEG for high resolution color still-image, and MPEG for full-motion video. These multimedia data
require vast amounts of storage, processing time, and transmission bandwidth. The standardization efforts can be viewed as selecting the l.iv-i
combination of these three (actors, interoperability, and timing consideration. According to this view, the .standard for each component is
reviewed with frame of background, coding, compression, and current status. With the information from the survey study, a simulation sliin \ s
conducted to investigate the performance of LAN where multimedia data are transmitted. The simulation results show that the hiph resoh:! ton
image browsing activity in a LAN will burden the low speed LANs The adoption of compression chips or high speed LANs such a.- FDDI will
make such high bandwidth activities feasible.
20 DISTRIBUTION/AVAILABILITY OF ABSTRACT
Q UNCLASSIHED.'lJNllMITtL) £j SAMt AS KtPOKl MIC UStKS
21 ABSTRACT SECURITY CLASSIFICATION
22a NAME OF RESPONSIBLE INDIVIDUAL
Myung W. Suh22b TELEPHONE (Include Area code)
(408)646 2637
22c OFFICE SYMBOlAS/Su
DD FORM 1473, 84 MAR 83 APR edition may be used until exhausted
All other editions are obsolete
SECURlTr CLASSiMCATiCi-lOF 'rv^.
Approved for public release; distribution is unlimited.
Multimedia Electronic Mail: Standards
and
Performance Simulation
by
Choi, Nag Jung
Captain, Republic of Korea Army
B.S., Korea Military Academy, 1987
Submitted in partial fulfillment
of the requirements for the degree of
MASTER OF SCIENCE IN TELECOMMUNICATION SYSTEMS MANAGEMENT
from the
NAVAL POSTGRADUATE SCHOOL
March 1992
&
ABSTRACT
This thesis surveys the current multimedia electronic mail (e-mail) related standards.
The increasing demands for interoperability pushed the international standardization
organizations to develop standards for e-mail. The multimedia e-mail related standards include
X.400 and ODA. ODA supports formatted text, raster graphics, and geometric graphics. The
future standards will be able to support high bandwidth uses such as high-resolution color
still-image, full-motion video, voice, and audio. The future components of multimedia e-
mail include TIFF for raster graphics, CGM for geometric graphics, JPEG for high resolution
color still-image, and MPEG for full-motion video. These multimedia data require vast
amounts of storage, processing time, and transmission bandwidth. The standardization
efforts can be viewed as selecting the best combination of these three factors, interoperability,
and timing consideration. According to this view, the standard for each component is
reviewed with frame of background, coding, compression, and current status. With the
information from the survey study, a simulation study is conducted to investigate the
performance of LAN where multimedia data are transmitted. The simulation results show that
the high resolution image browsing activity in a LAN will burden the low speed LANs. The
adoption of compression chips or high speed LANs such as FDDI will make such high
bandwidth activities feasible.
111
1.1
TABLE OF CONTENTS
I. INTRODUCTION 1
A. BACKGROUND 1
B. OBJECTIVES 2
C. LITERATURE REVIEW AND METHODOLOGY 3
D. ORGANIZATION OF STUDY 3
II. MULTIMEDIA ELECTRONIC MAIL AND ITS RELATED STANDARDS . 5
A. OVERVIEW 5
B. INTERNATIONAL ORGANIZATIONS FOR MULTIMEDIA MAIL
STANDARDIZATION 6
1. International Organization for Standardization (ISO) 6
2. International Telegraph and Telephone Consultative Committee
(CCITT) 7
C. X.400: MESSAGE HANDLING SYSTEMS (MHS) 8
1. Overview 8
2. The structure of MHS 9
D. OFFICE DOCUMENT ARCHITECTURE (ODA) 11
1. Overview 11
2. Document profile 13
3. Document structure 14
IV
DUDLEY F ->,Y
4. Interchange format (ODIF and ODL) 16
5. Document processing 18
6. Content structure 19
III. ODA CONTENTS STRUCTURE AND ITS CODING 20
A. CHARACTER CONTENT 20
B. RASTER GRAPHICS CONTENT 22
1. Background 22
2. Coding 22
3. Compression 23
4. Current status 23
C. VECTOR GRAPHICS CONTENT 24
1. Background 24
2. Coding 26
3. Compression 29
4. Current status 30
D. STILL IMAGE 30
1. Background 30
2. Coding 33
3. Compression 36
4. Current status 37
E. MOTION VIDEO: MPEG (Moving Picture Experts Group) 37
1. Background 37
2. Coding 39
3. Compression 42
4. Current status 44
IV. MODELING OF MULTIMEDIA NETWORK 45
A. PROBLEM DEFINITION 45
B. SYSTEM DEFINITION AND MODEL FORMULATION 47
1. System definition 47
2. Simulation model 47
C. MODEL TRANSLATION 51
1. SIMAN 52
2. LANNET II.5 53
D. EXPERIMENTS 54
V. SIMULATION RESULTS AND ANALYSIS 56
A. MODEL VERIFICATION 56
B. SIMULATION RESULTS AND DISCUSSION 57
VI. CONCLUSION 65
VI
APPENDIX A. SIMAN LAN MODEL AND EXPERIMENTAL SOURCE
CODE 67
APPENDIX B. SIMULATION RESULTS 69
APPENDIX C. GLOSSARY 70
APPENDIX D. LIST OF STANDARDS 74
LIST OF REFERENCES 78
INITIAL DISTRIBUTION LIST 85
vn
LIST OF FIGURES
Figure 1. The MHS and ODA in GOSIP environment 2
Figure 2. A MHS message and a letter 9
Figure 3. The MHS model 10
Figure 4. The composition of ODA 13
Figure 5. Layout structure 16
Figure 6. Document Processing Model 19
Figure 7. CGM and its relation to graphics application environment 24
Figure 8. Structure of a CGM 28
Figure 9. Organization of photographic standardization 32
Figure 10. Processing step for baseline DCT model 34
Figure 11. Three types of pictures 41
Figure 12. MPEG Coding process 43
Figure 13. Graphical representation of modeling LAN 50
Figure 14. An example of LANNET II.5 user interface 54
Figure 15. Utilization results from LANNET II.5 58
Figure 16. Average delay results from LANNET II.5 59
Figure 17. Utilization results from SIMAN 60
Figure 18. Average delay results from SIMAN 61
vin
I. INTRODUCTION
A. BACKGROUND
The technologies in telecommunication changed the concepts about geographical
distances and boundaries of the world. The innovative new technologies are implemented
and being offered to the public in breathtaking speed. One of the new technologies is
the multimedia electronic mail (e-mail). As in the telephone network, the increase of
value which the user gets from the e-mail system is proportional to the increasing number
of subscribers. If one subscriber cannot communicate with other subscribers, the value
of the network will decrease. The interconnection and interoperation with others
motivated the standardization efforts, both nationally and internationally, backed by
governmental organizations and private industries.
Multimedia e-mail is an application of multimedia computing and
telecommunications. It combines various media (e.g., text, raster graphics, vector
graphics, still-images, full-motion video, spreadsheets, animations, voice, and audio)
into electronic document form to transmit to remote recipients. The potential usefulness
of multimedia e-mail is enormous. It will significantly enhance the individual's
communication skill and efficiency by selecting the most suitable form of media for
communicating ideas.
The need for standards on multimedia e-mail system is widely recognized. For
example, the U.S. government adopted the standards for multimedia e-mail as a part of
Government Open System Interconnection Protocol (GOSIP). The two major standards
1
~ X.400 [Message Handling System (MHS)] and Office Document Architecture (ODA)
— are a part of GOSIP [Ref. 2]. Figure 1 illustrates the position of MHS and ODA in
GOSIP environment.
Application
Layer
Presentation
Layer
Office Document Architecture
(ODA)
Message
Handling
Systems
(MHS)(cenr 1984)
File Transfer, Access and Management
(FTAM) (ISO 857 1-4)
Virtual Terminal
(VT) (ISO 9041)
Association Control Service Element
(ISO 8650)
Connection-oriented Presentation Protocol
(ISO 8623)
Session Layer
Transport Layer
Network Layer
Data Link Layer
Physical Layer
Figure 1. The MHS and ODA in GOSIP environment
B. OBJECTIVES
There are many studies on multimedia e-mail systems, but some are too broad to
show detailed pictures and some are too narrow to show the whole scene. The main
intention of this thesis is to provide the reader with a clear, current, and in-depth view
of multimedia e-mail system and their standardization environment by surveying its
standards organizations and related standards. In addition to the survey part, a basic
simulation analysis is done to investigate the performance of a Local Area Network
(LAN) where multimedia data are transmitted.
C. LITERATURE REVIEW AND METHODOLOGY
One of the difficulties inherent in studying multimedia e-mail is the dynamic
standards environment in which it operates. The development life cycle of current
technologies is getting short. These fast and continuously changing technologies and
standards required the author to focus on the most up-to-date information at the time of
writing. In the survey study, most of the information came from current technical
journals, and computer and communication related magazines. General multimedia
applications, multimedia e-mail, and MHS related studies can be found in [Ref. 1 - 17].
General ODA information came from [Ref. 18 - 24], and ODA components standards
information came from [Ref. 25 - 80].
In the simulation part, the subset of multimedia LAN is modeled and simulated
using both LAN-oriented simulation tool and a general simulation language. Also,
related studies [Ref. 81-89] are investigated to support the results.
D. ORGANIZATION OF STUDY
The main body of this thesis is composed of four chapters: multimedia e-mail and
its related standards, components standards, modeling of a multimedia LAN, and
analysis of the simulation results.
Chapter II reviews the major international organizations for standardization which
include International Organization for Standardization (ISO) and International Telegraph
and Telephone Consultative Committee (CCITT). These organizations' structure and the
procedure of standardization are briefly reviewed. The multimedia e-mail related
standards are described with an emphasis on Message Handling Systems (MHS) and
ODA.
Chapter III surveys the content structure of ODA which defines the format of the
message components of multimedia e-mail. Included are current components (e.g. , text,
raster graphics, and geometric graphics) and future components (still-image and full-
motion video).
Chapter IV models a subset of multimedia LAN based on information from
previous chapters. The model describes an image browsing activity in a token ring
network which has transmission speeds of 4, 16, or 100 Mbps. The model takes into
account the effect of multimedia data, e.g. , vast amount of transmission needs, and the
application of compression technologies.
Chapter V verifies, presents, and analyzes the simulation results. Two properties
of LAN performance are collected and presented. The results from different simulation
tools and language are compared to verify the model.
Chapter VI contains a summary and conclusions.
H. MULTIMEDIA ELECTRONIC MAIL AND ITS RELATED STANDARDS
A. OVERVIEW
Electronic mail (e-mail) is now widely recognized as providing an effective means
of asynchronous communication between computer users, bridging the gap between the
traditional paper and telephone-based forms of communication. But, the current e-mail
supports the message which only contains text form of information. Multimedia mail is
an extension of the text-based e-mail model to incorporate formatted text, voice, stereo
audio, raster and vector graphics, animations, and full-motion video. The power of
multimedia mail lies in its ability to improve communications and promote group
collaboration by conveying information in one or more media to one or more users, thus
allowing the communication of ideas and concepts in the most suitable medium or
combination of media.
Compared with text-only e-mail, multimedia mail needs more sophisticated
technologies and standards for creating, editing, transferring, and viewing messages.
The standards are critical for multimedia mail. Without a standard representation format,
users can never be certain that any multimedia messages they send will actually be
readable by its recipients [Ref. 10]. Few users will spend time to create multimedia
documents if they do not expect them to be readable.
This chapter describes the role international organizations play in the multimedia
mail area and overviews their major outcomes: Office Document Architecture (ODA) and
X.400-series recommendations. ODA defines the architecture of documents to be
exchanged as contents of multimedia mail, and X.400 represents protocols for e-mail
interchange. This paper will focus on ODA and its components rather than on X.400
standards since the former has more direct impacts on network traffic than the latter.
Chapter III will provide a detailed survey of ODA components such as TIFF (Tagged
Interchange File Format), CGM (Computer Graphics Metafile), JPEG (Joint
Photographic Expert Group), and MPEG (Moving Picture Expert Group).
B. INTERNATIONAL ORGANIZATIONS FOR MULTIMEDIA MAIL
STANDARDIZATION
1. International Organization for Standardization (ISO)
The International Organization for Standardization (ISO) is a voluntary, non-
treaty agency of the United Nations. The scope of studies in ISO is very broad and
includes such fields as agriculture, nuclear systems, fabrics, documents used in
commerce, library science, computer systems, and computer communication [Ref. 1].
The majority of the work in ISO is handled in the technical committees (TC) and their
subunits named subcommittees (SC) and working groups (WG). The development of an
ISO standard is a time-consuming process because many countries and steps are involved.
A brief summary of the process is provided here [Ref. 17]:
• A new proposal is assigned to a technical committee (TC) which produces a
document containing the technical features of the specification. This document is
called a working draft (WD).
• As a result of technical review and editorial refinements on the WD, a modified
document is prepared. At this time, the document becomes a draft proposal (DP).
• The TC and the Central Secretariat review the proposal and issue it as a draft
international standard (DIS). At this juncture, the major comments and
corrections have been taken into account, and the document is circulated for a
member body vote.
• The DIS is reviewed and voted upon by members, and eventually published as an
international standard (IS). At this point, the document is considered to be
technically and editorially correct. It is then published in French, English, and
Russian.
Technical Committee 97 (TC97) deals with information technology. Its
subcommittee 18 (SC18) is developing standards for e-mail system interconnection, or
text preparation and interchange. Working Group 3 (WG3) of SCI 8 is developing
standards for a message body format and is further along in this area than most other
standards organizations. WG 4 is developing standards for a message heading format
and for massaging-related protocols [Ref. 8].
2. International Telegraph and Telephone Consultative Committee (CCITT)
The International Telegraph and Telephone Consultative Committee is a
member of the International Telecommunications Union (ITU), a treaty organization
formed in 1965. CCITT sponsors a number of recommendations dealing primarily with
data communications networks, telephone switches, and digital systems and terminals.
Like other standards organizations, CCITT is also criticized for its lengthy process in
making standards available to the public. CCITT's usual process of standardization is
summarized as follows:
• Standards group sets up a broad framework from which the standard could be
developed.
• The standard is made as detailed as possible, given a time constraint.
• The standard might be issued with certain pieces unfinished (with the notation
"Further Study").
• The standard is then published.
• As it is used, the "gaps" are filled in, deficiencies are noted, and modifications
are made in the light of practical operating experience.
• Whole process takes four years. At the end of each process, CCITT publishes
updated series of recommendations. Those books are identified by color of covers.
The 1960 books were red; 1964, blue; 1968, white; 1972, green; 1976, orange;
1980, yellow; 1984, once again red; 1988, blue [Ref. 16].
The CCITT Study Group VII is responsible for the X-series
recommendations. CCITT' s Study Group VII (SG VII) began work on e-mail system
interconnection officially in March 1981. Its comprehensive work should culminate in
specifications for message handling services: namely, the submission, relaying, and
delivery protocols; the structure of user names; the format of the heading and body of
messages; and methods of inter-working with Teletex terminals [Ref. 8]. The X-series
recommendations are categorized according to the functions and services they provide
and are then further classified into the specific recommendations.
C. X.400: MESSAGE HANDLING SYSTEMS (MHS)
1. Overview
CCITT published the initial Message Handling Systems (MHS) standards in
1984, and improved and extended the original standards in 1988. They consist of nine
recommendations from X.400 to X.420.
The MHS standards define two types of services to support e-mail. First,
the Message Transfer Service (MTS) is responsible for application-independent delivery.
Its operations are centered on the envelop of the message. Second, the Interpersonal
Messaging (IPM) Service supports communications with existing telex and telematic
8
services and defines specific user interfaces with MHS. It is responsible for the contents
within the envelope. Figure 2 shows the correspondence between an MHS message and
a conventional mail.
^^^From Address
To Address
Envelope
Heading':
:
:
:
:
:
:Address::
:
:
;; Body of letter :j:•
Body
Sigfty
Figure 2. A MHS message and a letter
2. The structure of MHS
MHS consists of five elements: User Agent (UA), Message Transfer Agent
(MTA), Message store (MS), and Access Unit (AU). Figure 3 shows the relationship
among these elements.
The User Agent (UA) is responsible for directly interfacing with the end user.
It prepares, submits, and receives messages for the user. The services UA provides
User
User:,
Message Handling System (MHS)
UA Message Transfer System (MTS)
Routing and relaying
UA
UA s
Submission and delivery
UA: User Agent
MTA: Message Transfer Agent
MS: Message Store
AU: Access Units
User
User
User
Figure 3. The MHS model [Ref. 16]
are summarized as follows:
• Text editing and presentation services for end users
• User-friendly interaction
• Security, priority provision, delivery notification, and distribution of subsets of
documents.
UAs are grouped into classes by MHS based on the types of messages they can process.
These UAs are then called cooperating UAs.
The Message Transfer Agent (MTA) provides the routing and relaying of the
message. This function is responsible for the store-and-forward path, channel security,
10
and the actual message routing through the communications media. A collection of
MTAs is called the Message Transfer System (MTS).
The Message Store (MS) provides for the storage of messages and support
their submission and retrieval. The MS complements the UA for machines such as
personal computers and terminals that are not continuously available. The job of MS is
to provide storage that is continuously available.
The Access Unit (AU) supports connections to other types of communications
systems, such as telematic services, postal services, etc.
The MHS conveys three kinds of information: message, probes, and
reports. The message consists of the envelope which carries the information needed to
transfer the message and the content which is the information the originating UA wishes
delivered to one or more UAs.
The probe is an information object that only contains an envelope. It is
delivered to the MTAs serving the end users. It is used to determine if the message can
be delivered.
The report is a status indicator. It is used by the MTS to relate the progress
or outcome of a message's or probe's transmission to one or more potential users.
D. OFFICE DOCUMENT ARCHITECTURE (ODA)
1. Overview
ODA is an international standard to facilitate the exchange of multimedia
documents, as defined by ISO 8613 [Information Processing - Text and Office System-
Office Document Architecture] developed by ISO TC97/SC18. The development started
11
in 1982 and was finally standardized after seven-year efforts in 1988. But the process
is still going on to allow us to add more components to a document. Current standards
only allow formatted text, raster graphics, and geometric 1
graphics, but the future
standards will include video, voice, audio, animation, etc.
Currently, American Standard Code for Information Interchange (ASCII) is
the only vendor-independent and most widely used format for text data exchange, but
it is useless for exchanging information regarding the layout and the character attributes
such as holding, underlining, and italicizing. On the other hand, ODA supports the
interchange of documents which contain picture and tabular materials. ODA will
promote the communication of the intentions of the author with respect to content and
appearance in different presentation media by means of a common set of rules. ODA
is a complex standard because almost every program for processing text, graphics, and
images has its own format for describing and storing information.
The ODA standard is composed of seven parts as seen in Figure 4. Part I
overviews the overall framework of ODA and the remaining parts describe details of the
following issues:
• Document profile
• Document structure
• Interchange format
• Document processing
• Content format.
1 Vector graphics is another name for geometric graphics.
12
ODA(ISO 8613)
Overview of the architecture
Document architecture
Document and application
ODIF and ODL- Document interchange format
text
Vector graphics
Raster graphics
>
Documentcontents
-Text-CGM-TIFF
Figure 4. The composition of ODA
Each of these issues will be briefly discussed in this section.
2. Document profile
The document profile is sent as a header at the start of data stream. The
profile can be sent separately from the document itself to help the recipient decide
whether he can process the document (recipient with text-only screen may not view the
graphics portion of the document). If the receiving system is not able to process the
profile, the originator need not bother sending the attached document.
The document profile consists of a set of attributes at the higher level in the
document structure. The first part identifies the title, author and date the document was
created. The second part includes a set of attributes applied to the document as a whole
13
[Ref. 22]. The role of document profile in a document is to give an overall definition
of the document characteristics, specifying the following:
• Its title, version number, subject, summary, table of contents, etc.
• Its history: author's name, day of creation, modification, period of validity,
organization responsibility for the document.
• The form of document: formatted, processable, or formatted processable.
• Its internal characteristics: nature of the objects, a list of fonts, character sets and
character attributes used within the entire document [Ref. 19, 22].
3. Document structure
The document structure answers the question of how ODA sees and describes
a document. In ODA, a document is an amount of structured information that can be
interchanged as a unit. The overall document architecture is the collection of rules for
defining the structure and representation of document [Ref. 22].
ODA structure consists of logical structure and layout structure. The logical
structure associates the content of the document with a hierarchy of logical objects such
as summaries, titles, sections, paragraphs, figures, and tables. The layout structure
associates the same content with a hierarchy of layout objects such as page sets, pages,
frames, columns, blocks, etc. This separation provides more than two means of
interchange. The layout view is useful for interchange of information regarding a
document's appearance; the logical view is useful for exchanging content that may be
revised.
The logical structure divides the contents of a document as a function of the
semantic description desired by the author. In this way, the notions of chapter,
14
paragraph, summary, note, indent, table of content, etc, may appear [Ref. 19].
There are three types of logical objects ODA allows: basic, composite, and root
(document). Basic objects have no subordinates and they include the objects such as
date, subject, paragraph, graphic, image, and signature. The composites have
subordinates of basic objects and they include header, summary, figure, etc. The root
is the top level and it has subordinates of composite objects.
To present a document on paper or on a screen, the logical structure is no
longer necessary. The document must then be divided into physical areas taking into
consideration of characteristics of the rendition equipment (e.g., screen, printer) and
the author's intentions expressed by layout styles associated with the logical objects. The
logical tree is converted into a layout tree — a tree-like structure of objects directly
adapted to the imaging of the document. These objects are page sets, pages and objects
within a page: frames and blocks [Ref. 19]. The layout structure of a document is
illustrated in Figure 5.
The page set is a composite layout object that consists of one or more pages
and/or one or more subordinate page sets, which need to be identified as a group.
A page is a rectangular area that corresponds to a unit of the presentation
medium (e.g., printer, screen, plotter). It is the reference area used for positioning
and imaging the content of the document.
A frame is a rectangular layout area within a page or within a frame of a
higher hierarchical level. Frames define boundaries within a page for the layout of the
content.
15
MANDATORY OPTIONAL
DOCUMENT
PAGE SETS
PAGES
FRAMES
BLOCKS
Figure 5. Layout structure
A block is a rectangular area containing one or more content portions of the
same content architecture [Ref. 20].
4. Interchange format (ODIF and ODL)
ODA provides two formats for document interchange: Office Document
Interchange Format (ODIF) and Office Document Language (ODL). ODIF was the only
format for ODA document interchange (1984 standard) and later the second format,
ODL, was added (1988 standard).
Both interchange formats work with three types of documents [Ref. 19, 22]:
• Processable document: the result of the editing process, which will be represented
by logical objects whose content is revisable. It may be both edited and formatted.
It is the form for the exchange of revisable documents without including
16
instructions for presentation. The logical generic structure and the styles can
accompany the document during interchange.
• Formatted document: the result of the layout process, which will be represented
by layout objects whose content is not itself revisable. This allows the interchange
of the image of a finished document. The layout generic structure may sometimes
accompany the document to factorize repetitive entities.
• Formatted processable document: the result of the editing and layout processes.
For this category documents, there will be coexistence of both logical and layout
structures. This form allows documents to be presented as well as edited and
formatted. Contents will include marks which will identify commands added
during the layout process.
a. ODIF
The Office Document Interchange Format defines the coding of objects,
styles and contents of structure documents. The interchange format specifies two things:
First, the sequence in which descriptors and text units appear (e.g., logical structure
descriptors precede layout structure descriptors); second, the abstract syntax for
interchange, as defined by ISO 8824 (Specification of Abstract Syntax Notation One:
ASN.l)and CCITT X.419.
ODIF is a binary encoding of ODA documents. It is a machine-
readable format designed for use in OSI network [Ref. 22]. Within a data stream, the
information is ordered according to the rules of the standard. There are two sets of rules
for ODIF: [Ref. 22]
• Interchange format class A: This set of rules is used to exchange any type of ODAdocument: formatted, processable, or formatted processable.
• Interchange format class B: This set does not support the generic or specific
logical structures and is therefore only used to exchange formatted documents for
presentation on recipient's system. This format is equivalent to sending a file in
a page description language.
17
b. Office Document Language (ODL)
ODL is an Standard Generalized Markup Language (SGML)2
application for describing ODA documents in human-readable format. ODL is
technically equivalent to ODIF, but there are several significant difference.
• ODL can send a document which is divided into several files. In ODIF, a
document must be sent as one complete file
• ODL provides a way of retaining the entire structure of a document using SGMLsyntax. In ODIF, all entities related to a document are transmitted as part of the
itself such that there is no means of retaining the entire structure of a document.
• ODL transmits a document in human-readable clear-text format.
• ODL provides a more convenient way for publication industry and DoD which
already make use of SGML.
5. Document processing
The document processing includes three processes: editing, layout, and
image generation. The editing process is used to create or reprocess a document. The
author manipulates the logical objects making up his document and the various contents
associated with them. If a document conforms to a prerecorded document class, this is
represented by a logical generic document class. Such a document is described by a
logical generic structure which guides the editing process by requiring the author to
respect the characteristics of the class [Ref. 19]. The layout process converts the logical
structure of the document into a layout structure. Finally, the image generation process
2 SGML is ISO standard 8879 (1986). It is a document representation language for
describing information in a standard notation, or syntax. Currently, there are two public
application of SGML in widespread use: one developed by the Association of American
Publishers (AAP) for use in books and journals and one developed by the U.S.
Department of Defense (DoD) for use in technical publications.
18
(presentation process) enables the document to be represented in a language adapted to
the imaging environment. This process can compose the image of a page by positioning
the character and graphic areas correctly and transfer the bitmap image of the page to a
printer or screen. One can imagine translations of the layout structure into page
description languages used today on various printers (e.g., PostScript, DDL)[Ref. 19].
Figure 6 shows the three processes of editing, layout, and image generation.
LOGICAL
OBJECTDEFINTrONS
DOCUMENTCLASS
DESCRIPTION
LAYOUTOBJECT
DEFJNITONS
PREVIOUS
LOGICAL
STRUCTURE
PREVIOUS
CONTENT
LOGICAL
STRUCTURE
EDITING
CONTENTEDITINO
EDITING
PROCESS
EDITED
LOGICAL
STRUCTURE
EDITED
CONTENT
LAYOUT
STRUCTUREf-)—Cf STRUCTURE
GENERATION
CONTENTLAYOUT *
LAYOUTPROCESS
LAYOUT
LAID-OUT
CONTENT
IMAGE
GENERATION
IMAGE
OF
DOCUMENT
.STRL RACTI.-.'. "ROC 3SS TwOinii SO PROCESS
Figure 6. Document Processing Model [Ref. 20]
6. Content structure
The ODA document content includes characters, geometric graphics, and
bitmap graphics. These are the basic contents which ODA currently supports. Chapter
III will give more detailed discussions on each of these components.
19
m. ODA CONTENTS STRUCTURE AND ITS CODING
Until now, ODA allows formatted text, raster graphics, and vector graphics. But
the future standards will include still image, motion video, and audio. Current ODA
content standards approved by ISO are formatted text, TIFF (Tagged Image File
Format) for raster graphics which is also standard format for Group IV FAX, and CGM
for vector graphics.
This chapter will survey the general background and current status of each of these
standard as well as future considerations for still image and full-motion video.
A. CHARACTER CONTENT
While character content architecture is based on several ISO and CCITT standards
and recommendations3, its basic architecture is ISO 8613-6 [Information Processing -
Text and Office Systems-Office-Document Architecture (ODA) and Interchange Format -
Part 6: Character Content Architectures] and CCITT Recommendation T.416 (Blue
Book, 1988) [Open Document Architecture (ODA) and Interchange Format - Character
3. ISO 2022 (1986): Information processing - ISO 7-bit and 8-bit coded character
sets - Code extension techniques.
. ISO 6429 (1983): Information processing - ISO 7-bit and 8-bit coded character
sets - Additional control functions for character imaging devices.
. ISO 6937 (1983): Information processing - Coded character sets for text
communication.
CCITT Rec. T.50 (1988): International alphabet No. 5.
CCITT Rec. X.208 (1988): Specification of abstract syntax notation one (ASN. 1).
CCITT Rec. T.61 (1988): Character repertories and coded character sets for the
international Teletex service.
ISO 9541-5 (1991 DS): Information processing - Font and character information
interchange - Font attributes and character model.
20
Content Architectures]. Character content part of ODA consists of fourteen sections
dealing with character content formats, character positioning and imaging, attributes,
definitions, control functions, layout and imaging process, etc.
As in ODA document, there are three classes of character content format:
formatted, processable, and formatted processable. Formatted content is for layout and
imaging of that content, but not for editing. Processable content is content which has
not been laid out. Formatted processable content is content that is structured such that
it contains both the formatted content and the processable content as subset.
The positioning of a character is decided by directions, position point, character
baseline, and escape point. Start-aligned, end-aligned, centered, and justified provide
alignment of a block. To support international language, it support reversed
presentation direction.
Four groups of specifications apply to the imaging of graphic character elements.
Those are emphasis, font selection, subscript and superscript, and character
combinations. Methods of emphasis are weight (faint, normal intensity, and bold),
posture (not italicized, italicized), underlining (not underlined, underlined , doubly
underlined) . Blinking (steady, slowly blinking, rapidly blinking), image inversion
(positive image, BEBBfflS), and crossing-out (not crossed-out, crossed out). For
font selection, the document profile specify the font used and characteristics. ODA
character content architecture follows ISO font standard ISO 9541.SuPcrscriPl and subscript is
accomplished by partial line down (PLD), partial line up (PLU), line position backward
(VPB), and line position relative (VPR) control functions.
21
B. RASTER GRAPHICS CONTENT
1. Background
Raster graphics part of ODA is Tagged Image File Format (TIFF) which
supports high-resolution bit-mapped gray-scale and color images, with compression
optional. Its related standards are CCll'l recommendation T.4 (1988) [Standardization
of Group 3 Facsimile Apparatus for Document Transmission], T.6 (1988) [Facsimile
coding Schemes and Coding Control Functions for Group 4 Facsimile Apparatus] and
X.208 (1988) [Specification of Abstract Syntax Notation One (ASN.l)]. There are two
classes of raster graphics content architecture. One is formatted, which allows for
document content to be presented as intended by the originator. Another class is
formatted processable, which can be processed as well as be presented as intended by
the originator.
2. Coding
Raster graphics content uses following encoding schemes:
• Group 4 facsimile encoding scheme (CCITT Rec. T.6)
• Group 3 facsimile encoding scheme (CCITT Rec. T.4)
• bitmap encoding scheme
CCITT Rec. T.6 specifies the principle of the Group 4 coding scheme as
follows:
The coding scheme uses a two-dimensional line-by-line method in which the
position of each changing picture element on the current coding line is coded with
respect the position of a corresponding reference element situated on either the
coding line or the reference line which is immediately above the coding line. After
the coding line has been coded, it becomes the reference line for the next coding
line. The reference line for the first line in a page is an imaginary white line.
22
The coding line is referenced to previous line and run length is decided by three coding
modes (pass, vertical, and horizontal mode). Then, the run length is translated into
codeword.
The Group 3 apparatus uses one-dimensional run length coding scheme. Rec.
T.4 describes its coding scheme as follows:
A line of data is composed of series of variable length code words. Each code
work represents a run length of either all white or all black. White runs and runs
alternate. A total of 1728 picture elements represent one horizontal scan line of
215 length. The code words are of two types: Terminating code words and Mark-
up code words. Each run length is represented by either one Terminating code
word or one Mark-up code word followed by a Terminating code word.
Another coding scheme used in Group 3 facsimile is two-dimensional coding
which is optional and identical to Group 4 facsimile coding scheme.
Third encoding scheme is bitmap encoding in which each pixel or element
is mapped with one-to-one way. Each element is represented by a single bit which has
value '0' or ' 1' depending on the state of the element (pixel). The resulting array of bits
which represents a row is encoded by a string of octets. If the number of bits in each
row of the array is not a multiple of eight, then it is extended by the minimum number
of '0' bits such that the last bit aligns on an octet boundary.
3. Compression
Currently, both hardware and software compression provide maximum of
255:1 compression ratio on bitmapped file [Ref. 31].
4. Current status
TIFF file format is widely accepted and the majority of graphics applications
provides import and/or export of this file format both in PC and Apple computer.
23
C. VECTOR GRAPHICS CONTENT
1. Background
A graphical metafile is a mechanism for the capture, storage, and transport
of graphical information [Ref. 51]. It provides a data format for picture archiving, a
single standard interface to picture-generating devices, and the glue for unifying and
integrating distinct graphics applications and hardware/software systems in a distributed
computing environment. Figure 7 illustrates functions of Computer Graphics Metafile
(CGM) related to graphics application environment.
Application
Program/r
\
Device-independent
Graphics package
MetafileReader/
GeneratorA
Device
Driver
Metafile
Generator
J/
CGMm,
Figure 7. CGM and its relation to graphics application environment
The CGM is the ISO standard that defines the functionality and encoding for
pictures stored on computer systems [Ref. 55]. It is a component standard of ODA
24
documents consisting of text, raster graphics, and vector graphics. CGM for the
storage and transfer of picture-description information enables pictures to be recorded for
long-term storage, to be exchanged between graphics devices, systems and installations.
In October 1981, ISO TC97/SC5/WG2 [Computer Graphics] established a
metafile subgroup that was given the task of:
• developing a framework in which graphical metafiles can be studied, developed
and related to other standards;
• designing a metafile standard proposal and recommending whatever 'work items'
would be needed in this area to produce the desired standard.
The history of standardization is well described in [Ref. 43].
The metafile subgroup had its first meeting in 1882. The aim of this project was
to create a system-independent graphics metafile that can be used with a wide range
of systems and devices. The subgroup has based its work on experiences with
various metafiles and Graphics Kernel System (GKS) and its metafile. However,
the greatest impact has come from the metafile group of the American National
Standards Institute (ANSI) X3H33. The group has been developing a national U.S.
standard for a metafile that originally called "Virtual Device Metafile VDM"[ANSI82]. Meanwhile, the ISO metafile project has been names "Computer
Graphics - Metafile for the Storage and Transfer of Picture Description Information
(CGM)" . The CGM is a basic metafile containing static consistent pictures without
a picture structure and without a dynamic picture change facility. The CGM is
defined in a multi-part standard (ISO 8632/1 to IS08632/4, currently 1985 at DIS
stage [IS085c]). Part 1 of the standard defines the CGM functionality, Part 2
contains a character coding based on ISO 2022 code extension techniques, and
Part 3 defines a binary coding. In part 4, a clear text coding which can be
written, read, and edited in the same way as plain text is specified.
The CGM standard was published as an ISO standard in 1987. In the United
States, American National Standard Institute (ANSI) published it as ANSI X3. 122-1986.
The U.S. government adopted the CGM as a Federal Information Processing Standard
(FIPS) 128 [Computer Graphics Metafile (CGM)] in March 1987, and the Department
of Defence (DoD) published a military specification MIL-D-28003 [Digital
25
Representation for Communication of Illustration Data: CGM Application Profile] in
December 1988. Also, it has been given a considerable push towards very wide
acceptance by its adoption within the DoD Computer-Aided Acquisition and Logistics
Support (CALS) initiative.
2. Coding
Pictures are described in the CGM standard as a collection of elements of
different kinds, representing, for example, primitives, attributes, and control
information. CGM is a static picture capture metafile, which means it contains no
elements with dynamic effects on partially defined picture.
The CGM standard consists of four parts; Part 1: elements and structure,
Part 2: character encoding, Part 3: binary encoding, and Part 4: clear text encoding.
• Part 1 is a functional specification. All standardized elements are identified, their
parameterizations are described, and their meanings are defined.
• Part 2 defines a character encoding of graphics metafiles. Opcode and parameter
data are encoded by characters from ASCII (ISO 2022) character set. The
resulting encoded data consists of printable characters. The encoding is compact
and can be transmitted directly through standard character-oriented communications
services.
• Part 3 defines a binary encoding of graphics metafiles. It is intended for
applications in which speed of generation and speed of transmission are most
important. The format for encoded data are either chosen for their similarity to
data formats in computers, or designed for fast decoding and processing.
• Part 4 defines clear text encoding of graphics metafiles. It is human readable
format. It is transmissible with standard character-oriented services, but is not
very compact and relatively slow to generate and interpret. An important feature
is its ease of comprehension and manipulation using standard text editors. It has
a format-free notation, comparable to modern, high-level programming language
syntax.
26
A CGM is structured as a series of levels: metafile level, picture level, and
picture body level. A metafile consists of a metafile descriptor and several picture
descriptions. The metafile descriptor contains information valid for the whole metafile -
- version of a metafile, metafile description. Picture descriptions are self-contained.
Pictures are bounded by picture-by-picture delimiters. A picture starts with a picture
descriptor that contains information about how this particular picture is stored and
information about scaling for specific device. Picture body contains the definition of the
picture, describing the graphical content of the picture. The graphical primitive
elements defines the geometric objects that make up the picture ~ lines, text, circles,
curves, etc. CGM contains a rich set of primitives for lines, markers, filled areas,
text, and generalized raster function. The structure of CGM is illustrated in Figure 8.
For color attributes, CGM provides two modes of color specifications -
indexed and direct. In indexed mode, color is selected by pointing to a color table
located at each workstation. For each COLOR INDEX value, the color table contains
an entry which specifies color in terms of Red/Green/Blue (RGB) intensities in the range
[0, 1]. In direct mode, the parameters of the color-specification elements are RGB
triples. Initially, CGM only supported RGB color specification. But in later
amendment (amendment 3), CGM was enhanced to support CYMK
(Cyan/Yellow/Magenta/Black) and CIE (Commission Internationale L'Eclairage) color
models.
CGM specifies three ways of encodings, i.e. character, binary, and clear-
text, for the functional description. The character encoding uses American Standard
27
BEGINMETAFILE
Metafile
DescriptorPicture
END
METAFILE Metafile
BEGIN
PICTUREPicture
DescriptorBody
ENDPICTURE
Picture
BEGINPICTUREBODY
Control, Primitive and Attributes ElementsPicture
Body
Figure 8. Structure of a CGM [Ref. 56]
Code for Information Interchange (ASCII) printing characters, the binary encoding uses
8-bit bytes, and the clear-text encoding is human-readable. A metafile in one encoding
can be translated to another encoding without loss of information. A number of graphics
packages can read in a CGM in one encoding, and output it in another. There are also
graphics utilities that can convert one encoding to another. The methods, advantages
and disadvantages of each encoding methods are summarized in Table 1
.
Table 1. CGM ENCODING
character encoding binary encoding clean-text encoding
encoding
method
ASCII printing
characters
8-bit byte clear-text (human
readable)
28
advantages . concise
. can be
transmitted over
any network
. easier
. quicker
. fairly concise
. can be edited with
normal text editor
disadvantage . require more
processing time
. cause problems
for some networks
. not concise
3. Compression
The compression ofCGM has less significance than the compression of raster
graphics or full-motion video because the size of a CGM metafile is relatively small
compared to other components, and the compression ratio can vary with the encoding
methods. Arnold, Liapakis, Reynolds, and Vezirgiannis [Ref. 56] conducted
experiments on the efficient encoding of metafiles for transmission over network. To
reduce the overall length of a metafile, they tried reordering the elements, delaying of
the encoding of an attribute change until such time as a primitive element that uses the
attribute is encountered, and catchall shorthand for the encoding of a metafile element
list by setting it to DRAWING PLUS CONTROL SET. They devised, implemented,
and tested the compression techniques for the binary encoding. The result is that an
average of 9 % of space was saved.
Another consideration can be given to the compression of metafile itself.
Current data compression utilities4can give from roughly 2:1 to 3:1 compression on
clear-text files. In the cases of character and binary encoding, the range of compression
4 Reduce, Stacpack, PKZIP, ARJ221A, LHA213, and PAK251
29
ratio roughly varies between 1:1 and 3:1. If CGM specific compression technique is
devised, the ratio will be greater.
4. Current status
Currently, there are a number of ISO and ANSI graphics standard,
including Initial Graphic Exchange Specification(IGES), Graphical Kernel System
(GKS), GKS-3D(GKS-dimensional), CGM, and Programmers' Hierarchical Interactive
Graphics System (PHIGS). Among them, CGM gets wide a support from industry and
government sectors. The wide adoption of CGM as a graphics standard is demonstrated
by the following events:
• CGM was adopted as the picture-defining protocol of ODA.
• In March 1987, the U.S. government announced its adoption of CGM as the
Federal Information Processing Standard (FIPS 128) [Ref. 52].
• In December 1988, U.S. DoD published its military specification MIL-D-28003
[Digital representation for communication of illustration data: CGM application
profile.] [Ref. 48].
• In May 1991, the U.S. government starts the CGM test service to determine
whether the CGM format is both successful and a practical method of sharing
images across diverse computer platform.
On the other side, the CGM is revised to provide the needs of application
groups. The first amendment was done in 1990 and third amendment is in process.
D. STILL IMAGE
1. Background
Still-image is one of the components to be included in ODA. The problem
with still-image consists in its vast amount of data to transmit and store. For example,
30
a frame of CCIR (A-601) digital studio video signal, represented by 720 x 486 pixels
and 16 bits per pixel, amounts to approximately 5.6 Mbits in an uncompressed form.
Modern image compression technology offers a solution to this problem.
State-of-the-art technology can compress typical images by a factor of 10: 1 through 50:
1
without visibly affecting image quality [Ref. 61]. But compression only is not sufficient
for modern open system environment in which interoperability is the major concern. The
receiver is supposed not to worry about what compression technique and equipment
sender uses. A standard compression method is needed to ensure interoperability of
equipment from different manufacturers.
Formal effort for still-image coding started at the end of 1986. Experts from
both the CCITT and ISO groups met to form the Joint Photographic Experts Group
(JPEG). They tasked themselves with the selection of a high performance universal
compression technique [Ref. 58]. The relationship between CCITT, ISO, and JPEG
is illustrated in Figure 9.
The first image quality evaluation took place in June 1987 at the Copenhagen
Telephone Company (KTAS) research lab. Out of the twelve proposed techniques, ten
31
— Facsimile
CCITT SG VIITerminal Quncfcnabcj
And Piococos
CCITT SGI
Services
VideotexDAPA
Architecture
New Image
Communication
Group {NIC)
DocumentArchitecture
Phototelegraphy
Joint Photo-
graphic Expert
Group (JPEG)
Joint Bi -level
Image Group
Computer Graphics
Coding
Photographic
Audio Coding —
CCITT SG XVVideophone £ Video-
Conference Services
1— Teleconferencing
Moving Picture
Coding
ISO ffiC/JTCl/
SC2/WG8
Oxlod Rcpreseoaooe
Of Picture And
Audio Tnfi"ff lulling
Digital Audio and
Picture Architectsr
(DAPA)
Figure 9. Organization of photographic standardization [Ref. 58]
techniques were tested at the evaluation5
. The selection procedures are well described
in [Ref. 58] and [Ref. 59]. The finalists were the Adaptive Discrete Cosine Transform
(ADCT), the Adaptive Binary Arithmetic Coder (ABAC), and the Block Separated
Component Progressive Coding (BSPC). BSPC is based on GBTC and PCS. The
second selection process took place in January 1988 at the same place, and resulted in
DCTV (Discrete Cosine Transform with Vector Quantization)
ADCT (Adaptive Discrete Cosine Transform)
DCTD (Adaptive DCT and Differential Entropy Coding)
BCTF (16 x 16 DCT with Filtering)
BLT (Block List Transform)
PCS (Progressive Coding Scheme)
PRBN (Progressive Recursive Binary Arithmetic)
ABAC (DPCM Using Adaptive Binary Arithmetic)
GBTC (Generalized Block Truncation Coding)
CVQ (Component Vector Quantization)
32
the selection of ADTC for final refinement, with the goal of producing and ISO Draft
Proposal by February 1989 [Ref. 59].
2. Coding
The proposed standard contains the four modes of operation [Ref. 61]:
• Sequential encoding: Each image component is encoded in a single left-to-right,
top-to-bottom scan.
• Progressive encoding: The image is encoded in multiple scans for applications in
which transmission time is long, and the viewer prefers to watch the image build
up in multiple coarse-to-clear passes.
• Lossless encoding: The image is encoded to guarantee exact recovery of every
source image sample value (e.g., as required by medical applications).
• Hierarchical encoding: the image is encoded at multiple resolutions, so that lower-
resolution versions may be accesses without first having to decompress the image
at its full resolution.
Based on these modes of operation, different kinds of codecs
(encoder/decoder) are specified. All of these codecs are variations of the JPEG baseline
model to ensure the operation with various image formats across applications. For the
DCT sequential-mode codecs, the baseline model shows almost complete coding process.
For the DCT progressive-mode codecs, the image buffer exists prior to the entropy
coding step, so that an image can be stored and then parcelled out in multiple scan with
successively improving quality. For the hierarchical mode of operation, the baseline
model is used as a building block within a larger framework.
The JPEG baseline model consists of 3 steps - transformation, quantization,
and lossless entropy coding as shown in Figure 10. At the input to the encoder, source
image samples are grouped into 8x8 blocks and precessed by the Forward Discrete
33
C~~9 :::: 8x 8 Block
Original
^ image ^i Encoder(>
FDCT r=5 Quantizer s•
DCCodingZig-Zag
Sequence=3
Entropy
Coder
Transmission /
Decoder '
IDCT C=DequantizerDC Decoding
Zig-ZagSequence
C=Entropy
Decoder
vFDCT: Forward Discrete Cosine Transform< rtv
Decompressed:i
g IDCT: inverse Discrete Cosme transform
image
Figure 10. Processing step for baseline DCT model
Cosine Transform (FDCT). Following is the description of the DCT part of JPEG
codecs [Ref. 61].
The DCT is related to Discrete Fourier Transform (DFT). Each 8x8 block of
source image samples is effectively a 64-point discrete signal which is a function
of the two spatial dimensions x and y. The FDCT takes such a signal as its input
and decomposes it into 64 orthogonal basis signals. . . . The output of the
FDCT is the set of 64 basis-signal amplitudes or "DCT coefficients" whose values
are uniquely determined by the particular 64-point input signal. ... At the
decoder, the IDCT reverses this processing step. It takes the 64 DCT coefficients
(which at that point have been quantized) and reconstructs a 64-point output image
signal by summing the basis signals.
The following equations are the idealized mathematical definitions of the 8 x 8 FDCT
and IDCT:[Ref. 61]
The next step is quantization which quantizes the DCT coefficients to reduce
their magnitude and to increase the number of zero value coefficients. The quantization
34
7 7
F(u,v)=-C(m)C(v)[£, X,/(x,y)cos^
—
77—cos J 1
4 x=0 y=0 10 10
7 7
^ >k IrY^^r/ \/-/\r/ \ (2x+l)U7C (2y + l)V7t,M>)=tE£ C(n)C(v)F(a,v) cos-^—
—
'-—cosv 7 y—
]
4 M=0 v=0 Id lo
where C(u) = C(v) = — for u = v = 0; C(k) = C(v) = 1 otherwise
v/2
method used in JPEG is uniform midstep quantization, where the step size varies
according to the coefficient location and which color component is encoded. The
purpose of quantization is to achieve further compression by representing DCT
coefficients with no greater precision than is necessary to achieve the desired image
quality. It is the principle source of lossiness in DCT-based encoders[Ref. 61].
Followed by quantization, DC coding and zig-zag sequence is applied. This
step rearranges quantized DCT coefficients into a zig-zag pattern, with lowest
frequencies first and highest frequencies last. The zig-zag pattern is used to increase the
run length of zero coefficients found in the block.
Next, the DC and AC coefficients are losslessly encoded, both using
Huffman-style coding but keyed with different parameters. Huffman coding is a well
known means of reducing the number of bits needed to represent a data set without losing
any information and is easy to implement in hardware. To compress data symbols, the
Huffman coder creates shorter codes for frequently occurring symbols and longer codes
for occasionally occurring symbols. The resulting data is passed to the host bus for
storage or transmission.
35
3. Compression
For color images with moderately complex scenes, all DCT-based modes of
operation typically produce the following levels of picture quality for the given range of
compression. (The units "bits/pixel" means the total number of bits in the compressed
image - including the chrominance components - divided by the number of samples in
the luminance component.) [Ref. 61]
• 0.25-0.5 bits/pixel: moderate to good quality, sufficient for some applications;
• 0.5-0.75 bits/pixel: good to very good quality, sufficient for many applications;
• 0.75-1.5 bits/pixel: excellent quality, sufficient for most applications;
• 1.5-2.0 bits/pixel: usually indistinguishable from the original, sufficient for the
most demanding applications.
A sample still-image with resolution of 720 x 486 and 16 bits per pixel -- 8
bits for luminance and 8 bit for chrominance components — will take up 5,598,720 bits
of storage without compression. Table 2 summarizes the relationship between
"bits/pixel" representation and compression ratio for this sample image.
Table 2. COMPRESSION RATIO FOR EACH BITS/PIXEL REPRESENTATION
bits/pixel Original image
size (bits)
Compressed image
size (bits)
Compression
ratio
0.25 5,598,720 87,480 64:1
0.50 174,960 32:1
0.75 262,440 21.3:1
1.00 349,920 16:1
1.50 524,880 10.7:1
2.00 699,840 8:1
16.0 5,598,720 1:1
36
4. Current status
Solutions for JPEG were offered first by C-Cube Microsystems Inc., San
Jose, Calif., then by LSI Logic Corp. , Milpitas, Calif [Ref. 67]. The first single-chip
image compression/decompression device CL550 from C-Cube Microsystems is 380,
000-transistor chip. CL550 can compress or decompress a full-page, 24-bit color, 300
dpi (dot per inch) image in less than one second, reducing the 25 megabyte original bit-
map image to under 1 megabyte with no visible degradation. A 640 x 480 resolution 24-
bit screen image can be compressed by a factor of ten in less than one-thirtieth of a
second, making real-time, full-motion compression feasible.
E. MOTION VIDEO: MPEG (Moving Picture Experts Group)
1. Background
Full-motion video will be the most demanding component of compound
document. The problem with full-motion video stems from its vast amount of data to be
processed and transmitted, which will severely strain CPU, storage, network
resources. For example, a frame of CCIR digital studio video (A-601) generates 5.6
megabits when uncompressed. For 30 frames per second video, the amount of data to
represent one second of motion comes up to 168 megabits. Liebhold and Hoffert [Ref.
72] compared the cost between voice and video transmission. Digital video running at
1 megabits per second bandwidth (compressed) would cost $1,000 per hour, compared
with voice running at 10 kilobits per second costing $10 per hour, assuming a coast-to-
coast connection.
37
Some of the standards for full-motion video are shown [Ref. 64, 65, 66,
70, 72, 79] in Table 3.
Table 3. FULL-MOTION VIDEO STANDARD
Name Organization Bandwidth Applications
H.261 CCITT 64 kbps x p(p=l ..30)
Video Phone, Video
Conference
MPEG ISO 1.5 Mbps CD-ROM, ISDN,
LAN
CD-I6 Industry 1.4 Mbps Consumer Electronics
CMTT/27 CCITT/CCIR 34 Mbps/45 Mbps Digital TV, HDTV
DVI8Intel 1.5 Mbps Desktop video
It is believed that MPEG will become the standard for the motion video
component of ODA document. MPEG standard is known by the name of the expert
group that started it: Moving Picture Experts Group which is part of the ISO-
IEC/JTC1/SC2/WG11 [Ref. 69]. Formal standardization efforts started in 1988 with the
goal of achieving a draft standard by 1990. After satisfying the requirements for the
standard, the selection process was begin. Seventeen companies or institutions
contributed or sponsored a proposal, and fourteen9different proposals were presented
6 Compact Disc-Interactive (CD-I)
7 Committee on Mix of Telephone and Television
8Digital Video Interactive (DVI)
9 AT&T (USA), Bellcore (USA), Intel (USA), GCT (Japan), C-Cube Microsystem
(USA), DEC (USA), France Telecom (France), Cost 211 Bis (EUR), IBM (USA), JVCCorp (Japan), Matsubisi EIC (Japan), Mitsubisi EC (Japan), NEC Corp (Japan), NTT(Japan), Philips (Netherlands), Sony Corp (Japan), Telenorma/U. Hannover (Germany)
38
and subjected to analysis and subjective testing. Those proposals were tested and
integrated into the Committee Draft in September 1990.
2. Coding
The industry- and sponsor-contributed proposals are representing the interests
of their own field— telecommunications, hardware, computer systems, consumer
electronics, and TV broadcasting. The coding method should be selected in such a way
that it can support a wide range of applications. For instance, video on digital storage
includes the following applications:
• Asymmetric Applications (one time compression, frequent decompressions):
electronic publishing, education and training, point of sales, videotext, games,
and entertainment (movies)
• Symmetric Applications (equal use of compression and decompression): electronic
publishing (production), video mail, videotelephone, and video conferencing.
Each application requires specific features in the coding algorithm. For
example, commercial video disc systems may need features of still frame, video-audio
synchronization, and fast forward/reverse search, but may not need the editability
feature. To fulfil requirements of various applications, MPEG defined a set of features
to be satisfied by its compression algorithm. These features are;
• Random access: Random access requires that a compressed video bit stream be
accessible in its middle and any frame of video be decoderable in limited amount
of time (1/2 second).
• Fast forward/reverse searches: It should be possible to scan a compressed bit
stream and, using the appropriate access points, display selected pictures to obtain
a fast forward or fast reverse effect.
• Reverse playback: Interactive application might require the video signal to play in
reverse.
39
• Audio-vidual synchronization: The video signal should be accurately synchronizable
to an associated audio source.
• Robustness to errors: The source coding scheme should be robust to any remaining
uncorrected errors; thus catastrophic behavior in the presence of errors should be
avoidable.
• Coding/decoding delay: The algorithm should perform well over the range of
acceptable delays and the delay is to be considered a parameter (e.g., video
telephone - maximum 150 millisecond; publishing application - maximum 1
second).
• Editability: It is desirable to be able to edit units of a short time duration.
• Format flexibility: The coding algorithm should handle wide range of formats
(width and height) and frame rate.
• Cost tradeoffs: The algorithm should be implemented with a small number of
chips, and the coding process could be performed in real time.
To satisfy both high compression ratio and the random access requirements,
MPEG uses a combination of intraframe and interframe coding methods. The method
used for intraframe coding is DCT and those for interframe are predictive coding and
interpolative coding. MPEG uses three types of pictures: intrapictures, predicted
pictures, and bidirectional pictures. Intrapicture provides the random access points and
reference points to both predicted pictures and bidirectional pictures. Figure 1 1 shows
the relationship between three types of pictures.
For interframe coding, MPEG uses motion-compensated prediction and
motion-compensated interpolation. Motion-compensated prediction is the most widely
used techniques that exploit the temporal redundancy of video signals [Ref. 69]. It
assumes that "locally" the current picture can be modeled as a translation of the picture
at some previous time. Motion-compensated interpolation is a technique that helps satisfy
40
1 2 3 4 5 6 7 8
I B B B P B^ B^ B.
I: Interpictures P: Predicted Pictures B: Interpolated Pictures
FunctionProvides access points
for random access
Reference for future
predicted pictureNot used for reference
Compression
RatioModerate High Highest
Reference picture NoneA past picture
(Intra- or Predicted)
Both a past and a future
picture (Intra- or Predicte
Figure 11. Three types of pictures
some of the application-dependent requirements since it improves random access and
reduces the effect of errors while at the same time contributing significantly to the image
quality.
The basic motion-compensation coding unit used in MPEG is a 16 x 16 block
(microblock) which has types of intra, forward predicted, backward predicted, and
average. The motion information associated with microblock is coded differently with
respect to the motion information present in the previous adjacent microblock. The
differential motion information is further coded by means of variable-length code to
provide a greater efficiency.
41
As in JPEG standard, MPEG uses 8x8 DCT (Discrete Cosine Transform)
to reduce spatial redundancy. Next, DCT coefficients are quantized. Because MPEG
has both intra-coded picture and differentially-coded picture, it uses different quantizer
for each block. While both quantizers have a constant step size, their step size around
zero is different. Quantizer for intracoded blocks has no deadzone (i.e., the region that
gets quantized to the level zero is smaller than a stepsize) while quantizers for
nonintrablocks have a large deadzone. Quantized coefficients are further compressed by
entropy coding step. To reduce the impact of the motion information on the total bit
rate, variable-length coding is used. The variable-length code associated with DCT
coefficients is a superset of the one used in CCITT recommendation H.261 to avoid
unnecessary costs when implementing both standards on a single chip [Ref. 69]. The
coding process is illustrated in Figure 12.
3. Compression
The bandwidth for transmission or storage of full-motion video depends on
the source video format and compression ratio. For example, NTSC-like picture
quality, assuming 480 x 270 pixels, 30 frames per second, and 16 bits per pixel, will
consume about 62 Mbps transmission bandwidth. With 50:1 compression, the
bandwidth required will be about 1.244 Mbps.
The premise of MPEG is that a video signal and its associated audio can be
compressed to a bit rate of about 1.5 Mbps with an acceptable quality. The quality of
video compressed with the MPEG algorithm at rate of about 1.2 Mbps has often been
compared to VHS recording. For most source materials, artifact-free renditions can be
42
8x8 Block
Motion
Compensation
Encoder
Variable-length
CodingFDCT Quantizer
Entropy
Coder
16 xl6 block
Transmission
Decoder
Motion
DSOOBfCflMOBB
VariaMe-lcsgth
DecodingIDCT ^jDequantizerfcd ^Jr
FDCT: Forward Discrete Cosine Transform
IDCT: Inverse Discrete Cosine Transform
Figure 12. MPEG Coding process
obtained; but for the most demanding material, it is at times necessary to trade
resolution for impairments [Ref. 71]. MPEG is not restricted to any given bandwidth.
In fact the baseline standard is specified to operate up to 2.1 Mbps.
4. Current status
The MPEG standard is still in draft form, but it has wide support, including
Apple, IBM, Sun, and Intel. C-Cube Microsystems became the first company to
publicly demonstrate an MPEG decoder chip [Ref. 78]. The MPEG decoder is based on
a programmable processor which enable changes in the standard by modifying the RAM-
based microcode. The processor supports two modes: MPEG and JVC Extended mode.
In MPEG mode, video can be compressed at rates of up to 200: 1 and resolutions of up
43
to 704 by 576 pixels. JVC Extended mode, which is being developed with Victory
Company of Japan Ltd., offers compression rates up to 50:1 [Ref. 79].
44
IV. MODELING OF MULTIMEDIA NETWORK
A. PROBLEM DEFINITION
Radding [Ref. 13] described one of the future military applications of multimedia
network. The setting is one of perspectives of the Department of Defence CALS
(Computer-Aided Acquisition and Logistic Support) initiative.
A jet aircraft technician peers into the bowels of a malfunctioning engine,
searching for the source of the problem. Finally, he spots it. Buried deep within
the engine is the troublesome part. He will have to replace it. A complicated
procedure, to say the least.
The technician goes to his high-powered workstation attached to a network and
calls up the information on the part and the replacement procedure. An image of
the part seated in the engine appears. In another window, an instructor
demonstrates the repair procedure in full-motion video while the technician listens
through the audio channel as the instructor explains the process. Diagrams pop up
to further clarify key points. In a text window, he reviews lists of necessary parts
and tools he will need to complete the repair.
Still confused about an irregularity in this situation, the technician presses the help
key and a real-time image of a live supervisor pops up in another window. Using
the attached microphone, the technician discusses the particular problem with the
supervisor, who directs more information onto the technician's screen. The
technician points a video camera at the part in question to show the supervisor the
specific situation.
Implementation of such a system is possible with today's technology, but the lack
of standards is the major barrier. The realization of importance of standardization led
the industry and the standard organizations to the standardization of multimedia network
and multimedia applications as we saw in Chapters II and III.
In addition to the standardization efforts, one of the problems with above scene
is the vast amount of data storage and transmission bandwidth. In particular, formatted
45
text, raster graphics, vector graphics, still image, and full-motion video will
significantly influence the performance of local area networks (LANs) and
telecommunication networks. The typical transmission bandwidth or storage
requirements for each component of multimedia e-mail are summarized in Table 4.
Table 4. STORAGE REQUIREMENTS FOR MULTIMEDIA E-MAILCOMPONENTS
Coding Original Size (a page, Compressed Size Compression
Format frame, or 1 second of (page, frame, or 1 Ratio
frames) second of frame)
Text 24-32kbits (clean text)
80-160kbits (formatted
text: WP5.1 file with
style)
1-2 k 2.0-3.0
Raster 200dpi: 4,036,608 297,769 13.56
Graphics (bi- 300dpi: 8,960,000 442,217 20.26
level) 400dpi: 16,146,432 593,226 27.22
CGM several 100 kbits to depends on coding 1-3
several Mbits (binary, character,
or clean text
coding)
JPEG (CCIR 5.6 Mbits/frame 112-560 10-50
A-601 format) kbits/frame
(multi-tone
color)
MPEG 62 Mbits/sec 1.5 Mbits/sec 50-150 (vary
(NTSC with frame by
quality) frame)
Current networks are able to support formatted text, bi-level raster graphics, and
some vector graphics effectively. But multi-tone color raster graphics, complex vector
graphics, and full-motion video can be a disastrous source of network bottleneck.
46
Therefore the design of multimedia applications must be preceded by a careful analysis
of network performance in relation to their bandwidth requirements.
There have been some studies on multimedia data transmission on LANs [Ref. 84 -
89]. Chen and Lu [Ref. 84] analyzed the integrated voice and data transmission on
CSMA/CD network. Mark and Lee [Ref. 85] proposed a dual-ring LAN for integrated
voice/video/data services that have different communication service requirements (i.e.,
synchronous and asynchronous). Deng, Chiew, and Huang [Ref. 88] investigated the
performance of token-ring LANs with multimedia file transfer.
Related to these studies, we develop a simulation model to analyze the performance
of multimedia LAN. The primary purpose of this simulation model is to examine the
effect of the multimedia data traffic, especially high-resolution still-images, on the
performance of LAN. Another consideration is that the adoption of current compression
technologies into the model for still-image compression.
B. SYSTEM DEFINITION AND MODEL FORMULATION
1. System definition
Within an office information system or military application as we saw in the
above scenario, an image browsing activity is an important part of LAN usage. An
image browsing activity may require transfer of a large volume of information in a short
period of time.
Currently, the LAN standards are categorized into three LAN types:
CSMA/CD bus, token bus, and token ring. CSMA/CD bus LANs are specified by
IEEE 802.3 standards and currently have five options: 10BASE5 (Ethernet), 10BASE2
47
(thin Ethernet), 1BASE5, 1BASET, and 10BROAD36. Token-ring LANs include
IEEE 802.5 4Mbps, IEEE 802.5 16Mbps, and ANSI's Fiber Distributed Data Interface
(FDDI) 100Mbps. Lastly, token-bus LANs are specified by IEEE 802.4 standards and
have a few options. We selected token-ring LANs for our simulation study because of
their support of a wide range of transmission speed (4, 16, and 100 Mbps), their
efficient operation at very high data rates, and the ability to readily accommodate fiber.
The token-ring network has stations actively coupled by unidirectional point-
to-point links. Access to the transmission medium is controlled by a deterministic
algorithm: a special bit pattern, called the token, circulates around the ring and gives
the station holding the token momentary control of the transmission channel. Several
service (transmission) disciplines have been proposed for token-ring networks. The most
widely used service disciplines are the following [Ref. 86]:
• Exhaustive service. A station transmits packets until its buffers become empty.
Only then it releases the token to its downstream neighbor.
• Gated service. A station transmits only the packets that are present in its buffers
when it captures the token. Then it releases the token to the next station.
• Limited services. A station transmits packets until either its buffers become empty
or a maximum number L of packets has been transmitted, whichever occurs first.
• Ordinary service. A station transmits packets up to one packet every time it
captures the token. The ordinary service discipline is in fact a special case of the
limited service with L = 1.
• Timed-Token Service (TTS). A token rotation timer (TRT) is used in each station
to ensure the duration of each complete cycle of the token as it is seen by this
particular station. The TRT is reset every time the token arrives at the station and
immediately starts timing the duration of the new token rotation.
48
Compared with the IEEE 802.5 token-ring protocol, FDDI has two major
differences in token handling:
• An FDDI station seizes a free token by absorption. On the contrary, an IEEE802.5 station seizes a free token only by flipping a bit.
• In FDDI, a station issues a free token as soon as it completes the transmission of
its frames, even if it has not yet begun to receive its own transmission. In IEEE802.5, a station releases a free token only when it has received its owntransmission.
Another difference is that FDDI defines two types of traffic: synchronous and
asynchronous. For synchronous traffic, FDDI allocates a constant bandwidth for each
station. But for asynchronous traffic, the access right to the network is regulated by the
parameter TTRT (Target Token Rotation Time) and depends on the traffic.
2. Simulation model
The simulation model includes a LAN and its stations. The types of LAN
considered are token-ring LANs with transmission speeds of 4, 16, and 100 Mbps. The
service (transmission) discipline selected is the exhaustive service with 262,080
bits/frame and frame overhead of 200 bits/frame. It is assumed that the 100 Mbps FDDI
follows the general token-ring protocols and the differences from IEEE 802.5 protocols
are ignored because the stations in the model do not require synchronous transmission.
The token passing time is set to zero because the delay to pass a token is relatively small
compared to high volume still-image data.
In our simulation model, we assumed that the LAN has only one type of
stations that generate high resolution graphics images to send to other stations. To
emphasize the LAN utilization and delay characteristics, the processing time for
49
preparation, buffering, retrieval, and storage of images will not be included in the
statistics collection. Each station generates high resolution image (CCIR A-601 format)
file (about 5.6 Mbits/image) and the time between generation follows the exponential
distribution with a mean of 60 seconds. We also assume that there is infinite size of
buffer at the destination, thus we do not expect any delay in this way. The number of
stations varies from 2 to 20 while other parameters stay the same. Figure 13 shows the
components of the model.
Figure 13. Graphical representation of modeling LAN
As for source of multimedia LAN traffic, we consider only one type of
source - still-image. There are several reasons for the use of only one type of traffic.
First, the text file is very small compared to high volume multimedia data. Second,
50
the bi-level raster graphics are a subset of still-image and their size is also relatively
small. Third, a CGM file can vary greatly in size and its transmission requirement
varies according to coding methods (e.g., binary, character, and clean-text coding);
but, in any way, its size is relatively small compared with a rasterized still-image file.
Fourth, the full-motion video requires synchronous transmission of data. When
compression technologies are introduced on full-motion video, there are complex
bandwidth requirements. These requirements and complexity go beyond the scope of this
study.
In conjunction with the fixed size image (5.6 Mbits/frame), the use of JPEG
still-image compression chip is considered. As we described in Chapter III, the JPEG
compression chip is available now and can achieve a compression ratio of 10:1 to 50:1
in one thirtieth of a second. The time for compression is not considered in this model.
We assumed that the compression ratio of a JPEG chips follows a normal distribution.
The distribution function is such that the output of the compression has a mean value of
336,000 bits/frame, lower bound of 112,000 bits/frame, and upper bound of 560,000
bits/frame. It is further assumed that the normal distribution has the standard deviation
of 100,000 so that the compression the ratio moderately spread out between lower bound
and upper bound.
C. MODEL TRANSLATION
The token ring access mechanism may be modeled by a single server representing
the rotation of the free token. The model is translated into different simulation languages
51
~ SIMAN and LANNET II. 5. The former is a general purpose simulation language and
the latter is a special purpose LAN simulation tool. We now discuss each translation.
1. SIMAN
SIMAN is a general purpose simulation language developed by Systems
Modeling Corp. for modeling discrete, continuous, and/or combined systems. SIMAN
is designed around a logical modeling framework in which the simulation problem is
segmented into a model component and an experiment component. The model describes
the physical elements of the system (LAN, stations, image files, etc.) and their logical
interrelationships. The experiment specifies the experimental conditions under which the
model is to run, including elements such as initial conditions, resource availability,
type of statistics collected, and length of a run.
The pseudo-code for the LAN model is;
CREATE arrivals (entity) as many as the number of stations
ASSIGN entity identification (attribute)
DELAY exponentially with mean of 60 seconds
ASSIGN time stamp (attribute)
ROUTE transfer the entity to a station
STATION at the entity's own station
QUEUE to await next token
SEIZE the channel
DELAY a quantum.
RELEASE the channel
TALLY the transmission delay
The variation of the above model is the case of a randomly varying file size.
In that case, an entity is given a file size by normal distribution with the mean of
336000. Based on that file size, the transmission time without any delay is calculated
52
to collect delay time statistics. If the data is less than the data bits per frame, the
transmission time is calculated based on LAN speed.
_, ... databits + frameoverhead .
Transmisswntime= - secondLANspeed(Mbps)
The source code for the model is shown in Appendix A.
2. LANNET H.5
The LANNET II.5 is a special purpose simulation package developed by
CACI Product Company for simulation of LANs. It is a design tool which takes a user-
specified LAN description and provides measures of utilization, conflicts, delays and
response times.
In LANNET II. 5, it is not necessary to program the model because it already
contains the LAN parameters and characteristics. The graphical interface gives the user
easy implementation of LAN simulation. The LANNET II. 5 package consists of four
parts;
• The LANGIN - Used to describe the LAN to be modeled.
• SIMLAN - The LAN simulation engine.
• LANPLOT - Used to plot/graph simulation statistics.
• LANAN - Post-processed LAN animation.
The LAN to be simulated is described using LANNET II. 5 building blocks
of LAN, STATION, GATEWAY, ROUTE, and SDF (Statistical Distribution
Function). Each of these building blocks has a series of attributes whose values are
supplied by the user. The LANGIN allows a user to build a LAN description using
53
LANNET II. 5 building blocks. An example of LANGIN graphical user interface is
shown in Figure 14.
OK Ucrify Nancs+
EXPERIMENTAL TOKEN RING
LAN Type LAN Connection List
LAN Inplenentat ions GRAPHICS U0RKSTATI0N o
<§> IEEE 802.5 4Mb
O IEEE 802.5 loflb
o USER DEFINED TOKEN RING LAN
ADD STflTIOM ADD GATEUAV
Figure 14. An example of LANNET II.5 user interface
D. EXPERIMENTS
The experiment part of a simulation study generates the information needed to
fulfill the study's goal. To get enough information, a series of experiments are
performed using both SIMAN and LANNET II. 5. Even though LANNET II.5 provides
very convenient ways of constructing and simulating the LANs, it lacks the flexibility
of enabling customized and detailed modeling of LANs. It supports only one service
discipline, exhaustive service, in token ring LANs. These are some of the reasons why
a general simulation language is used for the model. The experiments include two parts:
54
the case of a fixed-size image file and the case of randomly varying image file size. In
both cases, three LAN types -- IEEE 802.5 4 Mbps, IEEE 802.5 16 Mbps, and FDDI
token ring LANs — are considered. For each of these three LAN types, ten experiments
are carried out by varying the number of stations from 2 to 20 by 2. Thus, sixty
experiments are conducted in total. In SIMAN, each experiment has 10 replications
without initialization to collect confidence interval for minimizing random variation.
The duration of the simulation is set to 4 hours to cover unstable initialization and
real situations. The graphical analysis of simulation output indicates that LAN utilization
stabilizes after 1 hour of simulation time, so the selection of the 4 hour length of
simulation time appears reasonable.
55
V. SIMULATION RESULTS AND ANALYSIS
A. MODEL VERIFICATION
Verification is to determine that a simulation computer program performs as
intended, i.e, debugging the computer model into a correctly working program [Ref.
90]. In this thesis, two verification methods are used to ensure the model works
properly. First, built-in trace and animation functions of both LANNET II.5 and
SIMAN are used for the purpose of finding logical errors. Second, to make sure that
the internal parameters and logic ofLANNET II. 5 are correct, LANNET II. 5 simulation
results are compared with the results of the SIMAN model.
Both LANNET II. 5 and SIMAN provide debugging tools, including interactive
trace and animation. In the trace, the state of the simulated system (e.g., the contents
of the event list, the state variables, statistical counters) is printed out on a file or
screen just after each event occurs and compared with manual calculations to see if the
program is operating as intended. The trace feature in LANNET II. 5 provides the
reports regarding activity, gateway, LAN, message, and station. The trace feature
in SIMAN enables examining in detail the movement of entities through the system. It
also gives a summary of the action taken at each block through which an entity has
passed.
In addition to the trace feature, animation is an effective way of validating the
model. An animation can immediately expose most logical errors. LANNET II.
5
provides a LAN animation tool which provides graphical LAN animation using
56
simulation results from SIMLAN. This tool makes it possible to actually see the
modeled LAN in operation, and trace messages describing simulation activity may be
displayed concurrently with the animation. The location and color of each construct is
automatically created from the LAN description. SIMAN has animation tools named
CINEMA. They provide a graphical interface that helps users in drawing layouts of the
modeled system and linking the model and graphical objects.
Using both trace and animation features, any logical error is identified and
corrected. After the logic verification process, the results from both LANNET II. 5 and
SIMAN are compared. The comparison confirms that the model is correct.
B. SIMULATION RESULTS AND DISCUSSION
Two different properties of LAN performance are observed. The utilization is
calculated as percentage of time a LAN is busy during the simulation run. The time
required for an image file to move from a source station to a destination station in a
LAN is defined as the total transfer time in this model.
The delay time is the difference between the transfer time and the transmission time
which is calculated by dividing the size of the image file by the LAN transmission speed.
A ring network provides a single shared path for all data. Hence, as traffic increases
from the attached stations, transfer delay increases. The transfer delay is the most
relevant for asynchronous traffic.
Figure 15 through Figure 18 show the simulation results of the mean LAN
utilization and delay with the varying number of stations ranging from 2 to 20.
57
50.0%
40.0%
(a) Fixed file size case
30.0% —
20.0%
10.0% —
0.0%
3.0%
2.5%
2.0%
6 8 10 12 14
Number of Stations
(b) Random file size case
16 18 20
1.5% —
1.0%
0.5%
0.0%
8 10 12 14
Number of Stations
16 18 20
Figure 15. Utilization results from LANNET II.5
58
microsecond
600000 —(a) Fixed file size case
500000 —
400000
300000
200000
100000
4 Mbps
16 Mbps
100 Mbps
microsecond
1500 -
1250 —
1000
750
500
250 —
6 8 10 12 14
Number of Stations
(b) Random file size case
8 10 12 14
Number of Stations
16
16
18 20
18 20
Figure 16. Average delay results from LANNET II.5
59
50.0%
40.0% —
30.0%
20.0% —
10.0%
0.0%
3.0%
2.0%
1.0%
0.0%
(a) Fixed file size case
-*— 4 Mbps-o— 16 Mbps-— 100 Mbps+ LOWER* UPPER
6 8 10 12 14
Number of Stations
(b) Random file size case
16
-*— 4 Mbps-<^- 16 Mbps-— 100 Mbps+ LOWER+ UPPER
18 20
8 10 12 14 16 18 20
Number of Stations
Figure 17. Utilization results from SIMAN
60
microsecond600000
500000
400000 —
(a) Fixed file size case
300000
200000
100000
—<>
microsecond1500
1250
1000 —
750
500
250
t -^r^^r~t~^T^T i
8 10 12 14 16 18 20
Number of Staitions
(b) Random file size case
8 10 12 14 16
Number of Stations
18 20
Figure 18. Average delay results from SEMAN
61
The confidence interval is an indication of the reliability of the estimator of the
mean. With 10 experiments for each case, 95% confidence intervals are calculated as
follows:
|jl = [x-h, x+h]
where
x: point estimator of the mean(sample mean)
h = h, 0.05/2 *®
The sample mean, x, and the sample variance, s2
, of x and x from the 10
observations are as follows:
x = > — , s\x) = > , s\x) = —
—
T 10t
9 10
As we see in Figure 17 and Figure 18, the wide interval is resulted from the large
underlying variability of the exponential distribution for the arrival process. It is
reasonable that the confidence interval for the average delay of randomly varying file size
case is wider than the fixed file size case, because the former has more random factors.
In the fixed file size case, the utilization grows linearly with the increasing number
of stations. The utilization for 4 Mbps LAN is approximately 45% when the number of
stations reaches 20. The 45% utilization rate is very high, possibly causing a major
network bottleneck. In the case of variable file size, where it is assumed the JPEG
compression chip is utilized, the utilization goes down to about 3 % when the number
of stations reaches 20. Comparing two situation, . i.e., 45% utilization in non-
compression situation and 3% utilization in compression situation with number of
62
stations 20, it justifies the adoption of compression chip which compresses image files
with a 10 to 50 compression ratio. Considering other LAN traffics (e.g, file retrieving,
messages, etc.), the use of compression chip is highly recommended.
Since the maximum delay measure can be changed by simulation length, the
maximum delay results are not presented. But with the given simulation length of 4
hours, the maximum delay ranges from 1.28 second at 2 stations to 6.2 seconds at 20
stations in 4 Mbps token ring.
In the 16 Mbps token ring LAN case, the performance is acceptable in both
utilization and average delay at given LAN load. Utilization of 12% and an average
delay of 300 microseconds at maximum load is reasonably acceptable in image browsing
activity. Therefore, it is concluded that the 16 Mbps LAN can provide acceptable
performance under certain situations (e.g., no other traffics, maximum of 20 stations,
mean job request interval of 60 seconds, and job size of 5.6 Mbits). From another point
of view, if we consider the multimedia network in which synchronous data (e.g.,
packetized voice and video) and asynchronous data are transmitted, the 16 Mbps
bandwidth and token-passing protocol may not satisfy the requirements. The
requirements of synchronous data are guaranteed bandwidth and delay. Current collision
and token- passing type LAN protocols cannot support guaranteed delay for synchronous
data. Currently, another LAN protocol is being developed by IEEE. The IEEE 802.9
LAN standard is standard for Integrated Voice and Data (IVD) LANs. The specification
defines two channels: a packet channel for data applications, and a circuit channel for
voice and digital video.
63
The FDDI token ring is the most promising candidate for the multimedia LAN.
With dramatically increased bandwidth, FDDI meets the requirements for both
traditional LAN applications and multimedia applications. The timed-token protocol of
FDDI provides guaranteed delay for synchronous data transmission even at high network
load, through it cannot guarantee response time smaller than the round trip delay.
To better support high-volume synchronous traffic, Mark and Lee [Ref. 85]
proposed a dual-ring LAN which supports real-time synchronous voice and video
services, and multilevel priority asynchronous data traffic types using a packet-switching
operation. FDDI-II has also been proposed. Martini and Thomas [Ref. 89] proposed
FDDI-II which includes circuit switching capability in addition to the packet switching
capability of basic FDDI.
64
VI. CONCLUSION
Multimedia e-mail will be the next step of developments in exchanging information
among remote users of the computer network. All the society stands to benefit from the
successful implementation of multimedia e-mail. The two barriers hindering the wide
use of multimedia e-mail are its enormous bandwidth requirements and the lack of
standards for exchange format. The bandwidth restrictions are being resolved by
advanced computing and the deployment of ISDN, B-ISDN, and other high-speed
telecommunication networks. The other barrier, the lack of standards, is also being
resolved by international standards organizations' effort. The representative two
organizations, ISO and CCITT, are devoted to the standardization of multimedia e-mail,
alone or jointly. The resulting standards are X.400 and ODA. X.400 was standardized
in 1984 and revised in 1988; ODA was standardized in 1988 and is currently being
revised to allow more components. ODA supports formatted text, raster graphics, and
geometric graphics. The future standards will be able to support high bandwidth uses
such as high-resolution color still-image, full-motion video, voice, audio, and
animation.
For each component, there also have been individual and joint standardization
efforts. Currently, CGM, JPEG, and MPEG standards are being developed or revised
for geometric graphics, still-images, and full-motion videos, respectively. The
addition of these components will increase bandwidth requirements. The simulation
study shows that the high-resolution image browsing activity in a Local Area Network
65
(LAN) will burden the low-speed LANs. The adoption of compression chips or high-
speed LANs such as FDDI will make such high bandwidth activities feasible. In the case
of full-motion video, the synchronous transmission needs requires the development of
new protocols.
The combination of the wide bandwidth, compression technologies of source data,
and new protocol will facilitate the deployments of true multimedia e-mail and related
multimedia applications, and the results will enhance productivity of the future society.
66
APPENDIX A. SIMAN LAN MODEL AND EXPERIMENTAL SOURCE CODE
1. MODEL
BEGIN, Y, SIMLAN;
CREATE, No_of_Stations: 0, 1;
ASSIGN: X(l) = X(l) + 1;
ASSIGN: StationJD = X(l);
GOON DELAY: EXPO(60);
ASSIGN: Timeln = TNOW + Transfer_Time;
ROUTE: 0, StationJD;
STATION, 1;
Ql QUEUE, 1:DETACH;STATION, 2;
Q2 QUEUE, 2:DETACH;STATION, 3;
Q3 QUEUE, 3:DETACH;STATION, 4;
Q4 QUEUE, 4:DETACH;STATION, 5;
Q5 QUEUE, 5:DETACH;STATION, 6;
Q6 QUEUE, 6:DETACH;STATION, 7;
Q7 QUEUE, 7:DETACH;STATION, 8;
Q8 QUEUE, 8:DETACH;STATION, 9;
Q9 QUEUE, 9:DETACH;STATION, 10;
Q10 QUEUE, 10:DETACH;STATION, 11;
Qll QUEUE, 11:DETACH;STATION, 12;
Q12 QUEUE, 12:DETACH;STATION, 13;
Q13 QUEUE, 13:DETACH;
67
STATION, 14;
Q14 QUEUE, 14:DETACH;STATION, 15;
Q15 QUEUE, 15:DETACH;STATION, 16;
Q16 QUEUE, 16:DETACH;STATION, 17;
Q17 QUEUE, 17:DETACH;STATION, 18;
Q18 QUEUE, 18:DETACH;STATION, 19;
Q19 QUEUE, 19:DETACH;STATION, 20;
Q20 QUEUE, 20:DETACH;QPICK, CYC:
Ql: Q2: Q3: Q4: Q5: Q6: Q7: Q8: Q9: Q10:
Qll: Q12: Q13: Q14: Q15: Q16: Q17: Q18: Q19: Q20;SEIZE: Channel;
DELAY: Transfer_Time;
RELEASE: Channel;
TALLY: Transmission Delay, INT(Timeln): NEXT(GOON);END;
2. EXPERIMENTAL
BEGIN;PROJECT, SIMLAN, NAGJUNG CHOI;ATTRIBUTES: Timeln:
StationJD;
VARIABLES: No_of_Stations, 2:
Transfer_Time, 1.4011;
QUEUES: 20;
STATIONS: 20;
RESOURCES: Channel, 1
;
DSTAT:NR(Channel), Channel utilization;
TALLIES: Transmission Delay;
OUTPUTS: DAVG(Channel Utilization), "utill-02.dat":
TAVG(Transmission Delay), "dlyl-02.dat";
; Simulation time is set to 2 hours (7200 seconds) and 10 replications
; without initialization
REPLICATE, 10, 0, 7200, N;
END;
68
APPENDIX B. SIMULATION RESULTS
1. Fixed file size case
Numberof
Stations
Utilization Average delay (microsecond)
4 Mbps 16 Mbps 100 Mbps 4 Mbps 16 Mbps 100 Mbps
2 4.50% 1.15% 0.01% 15400 1076
4 8.92% 2.28% 0.02% 47720 2813
6 13.50% 3.46% 0.03% 91170 5606 1.375
8 18.10% 4.61% 0.04% 126000 7355 0.918
10 22.70% 5.79% 0.05% 174000 9149 0.651
12 27.20% 6.97% 0.06% 229000 11860 1.503
14 31.80% 8.13% 0.07% 289000 14470 0.5268
16 36.20% 9.30% 0.08% 359000 16790 0.5095
18 40.90% 10.50% 0.09% 433000 20050 1.91
20 45.40% 11.60% 0.10% 519000 21650 1.996
Random file size case (compression is applied)
Number of
Stations
Utilization Average delay (microsecond)
4 Mbps 16 Mbps 100 Mbps 4 Mbps 16Mbp 100 Mbps
2 0.27% 0.07% 0.01% 121
4 0.56% 0.14% 0.02% 197.9 9.295
6 0.85% 0.21% 0.03% 355.8 24.68 0.5373
8 1.12% 0.28% 0.05% 368.8 24.54 0.6168
10 1.41% 0.35% 0.06% 496.3 26.19 1.367
12 1.69% 0.42% 0.07% 671.3 41.15 1.474
14 1.97% 0.49% 0.08% 889.8 56.21 1.661
16 2.24% 0.56% 0.09% 1025 59.7 2.078
18 2.52% 0.63% 0.10% 1101 65.91 2.631
20 2.80% 0.70% 0.11% 1205 71.13 2.514
69
APPENDIX C. GLOSSARY
ABAC Adaptive Binary Arithmetic Coder
ADCT Adaptive Discrete Cosine Transform
ANSI American National Standards Institute
ASCII American Standard Code for Information Interchange
ASN. 1 Abstract Syntax Notation One
AU Access Unit
BCTF Block Cosine Transform with Filtering
BLT Block List Transform
BSPC Block Separated Component Progressive Coding
CALS Computer-Aided Acquisition and Logistics Support
CCIR International Radio Consultative Committee
CCITT International Telegraph and Telephone Consultative Committee
CD-I Compact Disc-Interactive
CD-ROM Compact Disc Read Only Memory
CGM Computer Graphics Metafile
CIE Commission Internationale L'Eclairage
CMTT Committee on Mix of Telephone and Television
CPU Central Processing Unit
CVQ Component Vector Quantization
70
CYMK Cyan/Yellow/Magenta/Black
DARPA Defence Advanced Research Projects Agency
DCTD Adaptive DCT and Differential Entropy Coding
DCTV Discrete Cosine Transform with Vector Quantization
DFT Discrete Fourier Transform
DIS Draft International Standard
DoD Department of Defence
DP Draft Proposal
DVI Digital Video Interface
FDCT Forward Discrete Cosine Transform
FDDI Fiber Distributed Data Interface
FIPS Federal Information Processing Standard
FTAM File Transfer, Access and Management
GBTC Generalized Block Truncation Coding
GKS Graphics Kernel Systems
GOSIP Government Open System Interconnection Protocol
HDTV High Definition Television
IDCT Inverse Discrete Cosine Transform
IEEE Institute of Electrical and Electronic Engineers
IGES Initial Graphic Exchange Specification
IPM Interpersonal Messaging
IS International Standard
71
ISDN Integrated Service Digital Network
ISO International Organization for Standardization
ITU International Telecommunications Union
IVD Integrated Voice and Data
JPEG Joint Photographic Experts Group
LAN Local Area Network
MHS Message Handling System
MPEG Moving Picture Experts Group
MS Message store
MTA Message Transfer Agent
MTS Message Transfer System
NTSC National Television Systems Committee
ODA Office Document Architecture
ODIF Office Document Interchange Format
ODL Office Document Language
PCS Progressive Coding Scheme
PHIGS Programmers' Hierarchical Interactive Graphics System
PRBN Progressive Recursive Binary Arithmetic
RGB Red/Green/Blue
SGML Standard Generalized Markup Language
TC Technical Committee
TIFF Tagged Image File Format
72
TRT Token Rotation Timer
TTS Timed-Token Service
UA User Agent
VDM Virtual Device Metafile
VT Virtual Terminal
WD Working Draft
WG Working Group
73
APPENDIX D. LIST OF STANDARDS
1. Standards for compound document format
a. International standards
ODA ISO 8613-1, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 1:
Introduction and General Principles (1988)
ISO 8613-2, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 2:
Document Structure
ISO 8613-4, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 4:
Document Profile
ISO 8613-5, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 5: Office
Document Interchange Format (ODIF)
ISO 8613-6, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 6:
Character Content Architectures
ISO 8613-7, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 7: Raster
Graphics Content Architectures
ISO 8613-8, Information Processing Systems - Text and Office Systems -
Office Document Architecture (ODA) and Interchange Format - Part 8:
Geometric Graphics Content Architectures
CCITT Recommendation T.410 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Overview
CCITT Recommendation T.411 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Introduction and General
Principles
74
CCITT Recommendation T.412 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Document Structure
CCITT Recommendation T.414 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Document Profile
CCITT Recommendation T.415 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Document Interchange Format
(ODIF)
CCITT Recommendation T.416 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Character Content
Architectures.
CCITT Recommendation T.417 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Raster Graphics Content
Architectures.
CCITT Recommendation T.418 (Blue Book, 1988) Open Document
Architecture (ODA) and Interchange Format - Geometric Graphics Content
Architectures.
SGML ISO 8879, Information Processing - Text and Office Systems - Standard
Generalized Markup Language (1986)
DSSSL ISO/IEC/ JTC 1/SC 18, Document Style Semantics and Specification
Language (Currently Draft form)
b. Industry standards
MO:DCA Mixed Object: Document Content Architecture (IBM)
RTF Rich Text Format (Microsoft)
CDA Compound Document Architecture (DEC)
DDIF Digital Document Interchange Format (DEC)
EPS Encapsulated PostScript
PCL Printer Control Language (Hewlett-Packard)
75
2. Character
ISO 9541 Information Processing - Font and Character Information Interchange
ISO 2022 Information Processing - ISO 7-bit and 8-bit Coded Character Sets -
Code Extension Techniques (1986)
ISO 6429 Information Processing - ISO 7-bit and 8-bit Coded Character Sets -
Additional Control Functions for Character Imaging Devices (1983)
ISO 6937 Information Processing - Coded Character Sets for Text Communication
(1983)
ASCII American Standard Code for Information Interchange
3. Raster graphics
CCITT Recommendation T.6, Facsimile Coding Schemes and Coding Control
Functions for Group 4 Facsimile Apparatus (1984)
4. Geometric graphics
CGM ISO 8632, Information Processing System - Computer Graphics - Metafile
for the Storage and Transfer of Picture Description Information (1987;
Amendment 1, 1990; Amendment 3, 1991)
GKS ISO 7942, Information processing system - Computer Graphics - Functional
Specification of the Graphical Kernel System (GKS, 1985)
PHIGS ISO 9592, Information Processing Systems - Computer Graphics -
Programmer's Hierarchical Interactive Graphics System (PHIGS, 1988)
IGES Initial Graphic Exchange Specification
5. Digital video
a. International standards
JPEG ISO 10918, JPEG Digital Compression and Coding of Continuous-Tone Still
Images (Draft Standard, 1991)
76
MPEG ISO 11172, Coding of Moving Pictures and Associated Audio (MPEG,Committee Draft, 1990)
H.261 CCITT Recommendation H.261, Video Codec for Audio Visual Services at
px64 kbits/s (1990)
b. Industry Standards
DVI Digital Video Interactive
CD-I Compact Disc-Interactive
6. Message Handling Systems (MHS)
MHS CCITT Recommendation X.400, (Blue Book, 1988), Message Handling
Systems: System Model-Service Elements.
CCITT Recommendation X.401, (Blue Book, 1988), Message Handling
Systems: Basic Service Elements and Optional User Facilities.
CCITT Recommendation X.408, (Blue Book, 1988), Message Handling
Systems: Encoded Information Type Conversion Rules.
CCITT Recommendation X.409, (Blue Book, 1988), Message Handling
Systems: Presentation Transfer Syntax and Notation.
CCITT Recommendation X.410, (Blue Book, 1988), Message Handling
Systems: Remote Operations and Reliable Transfer Server.
CCITT Recommendation X.411, (Blue Book, 1988), Message Handling
Systems: Message Transfer Layer.
CCITT Recommendation X.420, (Blue Book, 1988), Message Handling
Systems: Interpersonal Messaging User Agent Layer.
CCITT Recommendation X.430, (Blue Book, 1988), Message Handling
Systems: Access Protocol for Teletex Terminals.
77
LIST OF REFERENCES
1. Gerd Wallenstein, Setting Global Telecommunication Standards: the Stakes, the
Players and the Process, Artech House, Inc., Norwood, MA, 1990
2. FIPS PUB 146, GOSIP Version 2 OSI Architecture, 1988
3. Harry Forsdick, Multimedia Mail Meeting Notes, Network Working Group:RFC:910, Aug. 1984
4. M. Yamada, T. Miyasato, and K. Hasuike, "Mixed Mode Document Processing
System," IEEE GLOBECOM, Vol. 2, pp. 1171-1175, 1987
5. Katsuhiko Harota, Kouta Motumo, and Kenzo Nakajima, "Enhanced MessageHandling Systems," IEEE GLOBECOM, Vol. 2, pp. 1202-1206, 1987
6. Laura R. Pate, "Trends in Multimedia Applications and the Network Models to
Support Them," IEEE GLOBECOM, Vol. 1, pp. 317-321, 1990
7. G.D.Flinchbaugh, P.L Martinez, and D. S. Rouse, "Network Capabilities in
Support od Multimedia Applications," IEEE GLOBECOM, Vol. 1, pp. 322-326,
1990
8. David D. Redell and James E. White, "Interconnecting Electronic Mail Systems,"
IEEE Computer, Vol. 17, No. 9, pp. 55-63, Sept. 1983
9. Edward A. Fox, "Standards and the Emergence of Digital Multimedia Systems,"
Communications of the ACM, Vol. 34, No. 4, pp. 26-29, Apr. 1991
10. Nathaniel S. Borenstein, "Multimedia Electronic Mail: Will the Dream Become a
Reality?," Communications of the ACM, Vol. 34, No. 4, pp. 117-119, Apr. 1991
11. Jerry Cashin, "E-mail Standard Marks Progression Toward OSI; Industry
Embracing X.400 for Interoperability; the Battleship Galactica of E-Mail,"
Software Magazine, Vol. 10, No. 15, pp. 91-95, Dec. 1990
12. "Office redefined: a New Model for a New Generation," Patricia Seybold's Office
Computing Report, Vol. 13, No. 1, pp. 1-8, Jan. 1990
13. Alan Radding, "Staging an Entrance: Standards and Experience Will Put
Multimedia up in Lights," MIDRANGE Systems, Vol. 4, No. 18, pp. 24-28, Sept.
1991
14. "Who sets the Standards?," The Seybold Report on Publishing Systems, Vol. 18,
No. 13, pp. 5-12, Apr. 1989
78
15. Edward A. Fox, "Standards and the Emergence of Digital Multimedia Systems,"
Communications of the ACM, Vol. 34, No. 4, pp. 26-29, Apr. 1991
16. Uyless Black, The X Series Recommendations, McGraw-Hill, Inc., New York,
1991
17. Uyless Black, OSI: A Model for Computer Communications Standards, Prentice
Hall, New York, 1991
18. "DSSSL: an Alternative to ODA, " The Seybold Report on Publishing Systems, Vol.
19, No. 7, p. 18, Dec. 1989
19. Yves Chauvel and Philippe Maurice, "The Architecture of an Electronic
Document," IEEE GLOBECOM, Vol. 2, pp. 1166-1 177, 1987
20. W. Horok, "Office Document Architecture and Office Document Interchange
Formats: Current Status of International Standardization," IEEE Computer, Vol.
18, No. 10, pp. 50-60, Oct. 1985
21. Mark Walter, "Compound documents: Interchange and Integration," The Seybold
Report on Desktop Publishing, Vol. 5, N. 11, pp. 10-25, Jul. 1991
22. "A Capsule View of the Architecture. (Office Document Architecture)," The
Seybold Report on Publishing Systems, Vol. 19, No. 7, pp. 7-13, Dec. 1989
23. "ODA (Office Document Architecture): What is it? What is it Good for?," The
Seybold Report on Publishing Systems, Vol. 19, No. 7, pp. 3-5, Dec. 1989
24. Bob Brown and Ellen Messmer, "Vendors Rallying to Push Open DocumentStandard," Network World, Vol. 8, No. 13, p. 2, Apr. 1991
25. CCITT Recommendation T.416, Open Document Architecture (ODA) andInterchange Format - Character Content Architectures, 1988
26. Mark Walter, "The ISO Font Standard: Looking Beyond the Desktop," The
Seybold Report on Desktop Publishing, Vol. 4, No. 6, pp. 24-30, Feb. 1990
27. Barry Simon, "Squeeze Play," PC Magazine, Vol. 10, No. 17, pp. 291-300, Oct.
1991
28. "Multimedia Now Within Reach as Hardware is Made Available and Applications
Arrive," Computergram International, No. 1791, Oct. 1991
29. Dana Blankenhorn, "Comdex: More from the Sands Convention Center,"
Newsbytes, Oct. 1991
30. Cheryl J. Goldberg, "Designers Seek Standards, Formats to Resolve File-Exchange
Problems," MacWEEK, Vol. 5, No. 36, pp. 40-41, Oct. 1991
79
31. John McCormick, "Hardware Offers Massive Compression for Graphics Files,"
Newsbytes, Sept. 1991
32. Susan Glinert-Stevens, "Picture Packer: Pack a Peck of Pixels," PC Sources, Vol.
2, No. 8, p. 299, Aug. 1991
33. Ian Ashdown, "PCX Graphics," C Users Journal, Vol. 9, No. 8, pp. 89-86, Aug.1991
34. Shmuel T. Klein, Abraham Bookstein, and Scott Deer, "Storing Text Retrieval
Systems on CD-ROM: Compression and Encryption Considerations," ACMTransaction on Information Systems, Vol. 7, No. 3, pp. 230-245, Jul. 1989
35. Christos Chamzas and D. L. Duttwei, "Progressive Encoding of Facsimile ImagesUsing Edge Decomposition (PED)," IEEE GLOBECOM, Vol. 2, pp. 1011-1015,
1988
36. Yasuhiko Yasuda, Yasushiro Yamazaki, Takahiko Kamae, and Kazuo Kobayashi,
"Advances in FAX," Proceedings of the IEEE, Vol. 73, No. 4, pp. 706-730, Apr.
1985
37. Dennis Bodson, Stephen J. Urban, Alan R. Deutermann, and Charles E. Clarke,
"Measurement of Data Compression in Advanced Group 4 Facsimile Systems,"
Proceedings of the IEEE, Vol. 73, No. 4, pp. 731-739, Apr. 1985
38. FIPS PUB 149, General Aspects of Group 4 Facsimile Apparatus, National
Institute of Standards and Technology, Nov. 1988
39. CCITT Recommendation T.417, Open Document Architecture (ODA) andInterchange Format - Raster Graphics Content Architectures, 1988
40. CCITT Recommendation T.4, Standardization of Group 3 Facsimile ApparatusforDocument Transmission, 1988
41
.
CCITT Recommendation T.6, Facsimile Coding and Coding Control FunctionsforGroup 4 Facsimile Apparatus, 1988
42. Kenneth R. McConnel, Dennis Bodson, and Richard Schaphorst, FAX: Digital
Facsimile Technology & Applications, Artech House, Inc., Norwood, MA, 1989
43. Enderle Kansy Pfaff, Computer Graphics Programming, Springer-Verlag Berlin
Heidelberg, Germany, 1987
44. A. M. Mumford, "Computer Graphics Metafile Standard - an Update,"
Computer-Aided Design, Vol. 23, No. 4, pp. 303-305, May. 1991,
45. K. W. Brodlie, D. A. Duce, and F. R. A. Hopgood, "The New Graphical Kernel
System," Computer-Aided Design, Vol. 23, No. 4, pp. 312-319, May 1991
80
46. MIL-D-28003, CGM Application Profile; Digital Representation forCommunication of Illustration, 1988
47. Peter R. Bono, "A Survey of Graphics Standards and Their Role in Information
Interchange," IEEE Computer, Vol. 18, No. 10, pp. 63-75, Oct. 1985
48. Neil McManus, "DVI Takes QuickTime to Full Screen: New Video Board Offers
30-fps Movies," MacWEEK, Vol. 5, No. 36, p. 38, Oct. 1991
49. Michael Slater, "Intel Announces DVI Chip Set; DVI Chip Set to Support JPEGand Other Standards," Microprocessor Report, Vol. 4, No. 22, pp. 6-10, Nov.1990
50. Peter R. Wilson, "Standards: Past Tense and Future Perfect?," IEEE ComputerGraphics and Applications, pp. 44-48, Jan. 1991
51. Lofton Henderson, Margaret Journey, and Chris Osland, "The Computer Graphics
Metafile," IEEE Computer Graphics and Applications, pp. 24-32, Aug. 1986
52. FIPS PUB 128, Computer Graphics Metafile (CGM), Federal Information
Processing Standards Publication, Mar. 1987
53. Pichard F. Puk and John I. McConnel, "GKS-3D: A Three-Dimensional Extension
to the Graphical Kernel system," IEEE Computer Graphics and Applications, pp.
42-49, Aug. 1986
54. Thomas Powers, Andrea Frankel, and David Arnold, "The Computer Graphics
Virtual Device Interface," IEEE Computer Graphics and Applications, pp. 33-41,
Aug. 1986
55. A. M. Mumford, "Computer Graphics Metafile," Computer-Aided Design, Vol.
23, No. 4, pp. 303-305, May 1991
56. D. B. Arnold, S. T. Liapakis, G. J. Reynolds, and N. P. Vezirgiannis, "Practical
Considerations in Transporting Computer Graphics Metafile," Computer-Aided
Design, Vol. 23, No. 4, pp. 306-311, May 1991
57. Andy C. Hung, "Image Compression: The Emerging Standard for Color Images,"
IEEE Computing futures, pp. 20-29, Win. 1989
58. Graham P. Hudson, Hiroshi Yasuda, and Istvan Sebestyen, "The International
Standardization of a Still Picture Compression Technique," IEEE GLOBECOM, pp.1016-1021, Vol. 2, 1988
59. Gregory Wallace, Roy Vivian, and Henning Poulsen, "Subjective Testing for Still
Picture Compression Algorithms," IEEE GLOBECOM, Vol. 2, pp. 1022-1027,
1988
81
60. Alain Legar, Joan L. Mitchell, and Yasuhiro Yamazaki, "Still Picture CompressionAlgorithms evaluated for International Standardization," IEEE GLOBECOM, Vol.
2, pp. 1028-1032, 1988
61. Gregory K. Wallace, "The JPEG Still Picture Compression Standard,"
Communications of the ACM, Vol. 34, No. 4, pp. 30-44, Apr. 1991
62. Product Announcement, "Image Compression Reduced to Single Chip; First Single-
Chip JPEG Compression/Decompression Device. (C-Cube Microsystems' CL550),"
Microprocessor Report, Vol. 4, No. 4, pp. 1-3, Mar. 1990
63. Denise Caruso, "Is CD-I already a flop?," Digital Media: A Seybold Report, Vol. 1,
No. 2, pp. 3-6, Jul. 1991
64. T. Russell Hsing, "Motion Detection and Compression Coding for Motion Video
Coders: Technical Review and Comparison," IEEE GLOBECOM, Vol. 1, pp. 60-
64, 1987
65. Ronald Plompen, Yoshinori Hatori, Wilfried Geuen, Jacques Guichard, MarioGuglielmo, and Harald Brusewitz, "Motion Video Coding in CCITT SG XV - TheVideo Source Coding," IEEE GLOBECOM, Vol. 2, pp. 997-1004, 1988
66. Mike Carr, Jacques Guichard, Kiichi Matsuda, Ronald Plompen, and Joachim
Speidel, "Motion Video Coding in CCITT SGXV - The Video Multiplex and
Transmission Coding," IEEE GLOBECOM, Vol. 2, pp. 1005-1010, 1988
67. Peng H. Ang, Peter A. Ruetz, and David Auld, "Video Compression makes big
Gains," IEEE Spectrum, pp. 16-19, Oct. 1991
68. Dldier Le Gall, "MPEG: a Video Compression Standard for Multimedia
Applications," Communications ofthe ACM, Vol. 34, No. 4, pp. 46-58, Apr. 1991
69. Ming Liou, "Overview of the px64 kbit/s Video Coding Standard,"
Communications of the ACM, Vol. 34, No. 4, pp. 59-63, Apr. 1991
70. Frans Sijstermans and Jan Van der Meer, "CD-I Full-motion Video Encoding on
a Parallel Computer," Communications of the ACM, Vol. 34, No. 4, pp. 82-91,
Apr. 1991
71. Andrew Lippman, "Feature Sets for Interactive Images," Communications of the
ACM, Vol. 34, No. 4, pp. 92-101, Apr. 1991
72. Michael Liebhold and Eric M. Hoffert, "Toward an Open Environment for Digital
Video," Communications of the ACM, Vol. 34, No. 4, pp. 104-112, Apr. 1991
73. David Marshak S, "Multimedia: the Promise, the Players, the Issues," Patricia
Seybold's Office Computing Report, Vol. 13, No. 4, pp. 1-11, Apr. 1990
82
74. "C-Cube to supply MPEG chip for CD-I, " Microprocessor Report, Vol. 5, No. 12,
p. 17, June 1991
75. Denise Caruso and Andreas Bechtolsheim, "Digital Movies for the Masses,"
Digital Media: A Seybold Report, Vol. 1, No. 3, pp. 10-12, Aug. 1991
76. "Motorola and C-Cube announce MPEG Development for Phillips Systems,"
Multimedia Computing & Presentations, Vol. 3, No. 8, p. 7, June 1991
77. John Wharton, "Who Needs Multimedia," Microprocessor Report, Vol. 5, No. 11,
pp. 13-14, June 1991
78. "C-Cube Demonstrates MPEG Decoder Chip; Chip Supports Base-Level MPEGand JVC Extensions," Microprocessor Report, Vol. 5, No. 7, p. 8, Apr. 1991
79. Connie Guglielmo, "C-Cube Chip to Decompress Video; Processor Based onMPEG Standard," MacWEEK, Vol. 5, No. 13, p. 18, Apr. 1991
80. D. Marshak, "Multimedia and the Network," Patricia Seybold 's Network Monitor,
Vol. 5, No. 5, pp. 18-20, May 1990
81. Chain Ziegler and Gerald Weiss, "Multimedia Conferencing on Local AreaNetwork," IEEE Computer, Vol. 23, No. 9, pp. 52-61, Sep. 1990
82. Luis Orozco, Richard Vallee, Bilal Awada, and Nicolas Georganas,
"Communication Services for Real-Time Multimedia Office Application," IEEEGLOBECOM, Vol. 1, pp. 301-305, 1990
83. K. Sohraby and G. P. Austin, "ISDN Primary Rate Interface Impact onPerformance of Integrated Voice and Data on CSMA/CD Networks — AMeasurement and Simulation Study," IEEE GLOBECOM, Vol. 2, pp. 912-919,
1990
84. Wei-Hong Chen and Chung-Chin Lu, "CSMA/CD/TDMA: A DynamicCombination for Voice and Data Integration," IEEE INFOCOM, Vol. 3, pp. 842-
849, 1990
85. Jon W. Mark and Byoung-joon Lee, "A Dual-ring LAN for Integrated
Voice/Video/Data Services," IEEE INFOCOM, Vol. 3, pp. 850-857, 1990
86. Dennis Karvelas and A. Leon-Garcia, "A General Approach to the Delay Analysis
of Symmetric Token Ring Networks," IEEE INFOCOM, Vol. l,pp. 181-190, 1991
87. Robert H. Deng, Wendy C. L. Chiew, and Kuan-Tsae Huang, "Analysis of TokenRing Networks with Both Finite and Infinite Buffer Capacity Stations," IEEEINFOCOM, Vol. 1, pp. 202-210, 1991
83
88. Robert H. Deng, Wendy C. L. Chiew, and Kuan-Tsae Huang, "Token-Passing
Systems with Batch Arrivals and Their Application to Multimedia File Transfer
over Token-Ring LAN's," IEEE 1NFOCOM, Vol. 3, pp. 1041-1049, 1991
89. Peter Martini and Thomas Meuser, "Real-Time Traffic in FDDI-II Packet
Switching vs. Circuit Switching," IEEE 1NFOCOM, Vol. 3, pp. 1413-1420, 1991
90. Averill M. Law and W. David Kelton, Simulation Modeling & Analysis, 2nd ed.,
McGraw-Hill, Inc., New York, 1991
91. C. Dennis Pegden, Robert E. Shannon, and Randall P. Sadowski, Introduction to
Simulation Using SIMAN, McGraw-Hill, Inc., New York, 1990
92. LANNET 11.5 User's Manual, CACI Products Company, La Jolla, CA, 1990
84
INITIAL DISTRIBUTION LIST
No. Copies
1
.
Defence Technical Information Center 2
Cameron Station
Alexandria, VA 22304-6145
2. Library, Code 52 2
Naval Postgraduate School
Monterey, CA 93943-5000
3. Professor Dan C. Boger, Code AS/Bo 2
Department of Administrative Sciences
Naval Postgraduate School
Monterey, CA 93943-5000
4. Professor Myung W. Suh, Code AS/Su 1
Department of Administrative Sciences
Naval Postgraduate School
Monterey, CA 93943-5000
5. Professor Keebom Kang, Code AS/Kk 1
Department of Administrative Sciences
Naval Postgraduate School
Monterey, CA 93943-5000
6. Library 1
P.O. BOX 77, Gong-neung-Dong
Do-bong-Gu, Seoul, 132-240
Republic of Korea
7. Library 1
Army Headquarter, Bu-nam-Ri, Duma-MyunNon-san-Gun, Chung-nam-Do, 320-919
Republic of Korea
8. Choi, Nag Jung 5
105-8 Weon-cheon-Li, Baek-san-Myun
Bu-an-Gun, Jeon-buk-Do,
Republic of Korea
85
!
ThesisC448856 Choi
c# l Multimedia electronic
mail.
Thee is
C448856 Choi
c.l Multimedia electronicmail.