Advanced

Micro
Devices

Multimedia Over FDDI

Conference Paper Reprint

1992 IEEE, Reprinted with permission, from Proceedings of IEEE Local Computer Networks, Minneapolis, Sept. 13–15, 1992. 17386A
 Multimedia Over FDDI
Amit Shah1 , Don Staddon2, Izhak Rubin3, Aleksandar Ratkovic4

Advanced Micro Devices1, IBM2 , IRI Corp. and UCLA3 , IRI Corp. and UCLA4

stations over a total fiber length of 200 km with full system

Abstract performance.
The FDDI MAC supports multiple classes of service:
This paper focuses on issues in multimedia networking Synchronous, Asynchronous and Restricted Asynchronous.
over FDDI and simulation results demonstrating the The asynchronous class is used for normal, bursty data
feasibility of multimedia over FDDI. Previous work in this traffic. For more constant, isochronous traffic, the
field has concentrated on the asynchronous channel in synchronous channel can be used and it provides
FDDI. Other examinations of the FDDI synchronous guaranteed bandwidth to each station on each token
channel have applied bursty traffic to the synchronous rotation. The restricted asynchronous service provides a
channel. Multimedia traffic is typically stream oriented. mechanism for extended and protected dialog between a
Our simulations were focused on simulating typical limited number of stations. This feature is rarely used.
configurations with a range of applications and traffic There have been several publications on FDDI [10],
models. The advantages of FDDI as an integrated [11], [12], [13], [14], [15], [16] which explain the protocol
voice/video/ data network are demonstrated. in further detail.

1: Introduction 2: Multimedia
The paper is organized as follows. Multimedia is a term used to denote a set of
A brief discussion on the basics of FDDI and applications, products, and technologies [1]. We define it
multimedia is followed by an overview of the various as the use of multiple means to communicate information
multimedia applications and issues in networking of via a computer. Whereas computers today are mainly used
multimedia applications such as voice and video. The for textual data information presentation and storage,
importance of parameters such as network delay, intra- multimedia uses the computer for text, natural and animated
packet jitter, and bandwidth is highlighted. images, rendered graphics, auditory stimuli and realistic
The simulation models are discussed. The methodology sound. There are two key components to multimedia:
involved testing the effect of variations in TTRT, network multiple output media and interaction.
size, packet size, and offered load on the multimedia Television as we know it today exemplifies one of the
streams of voice and video, using the IRI Planyst‘ tool. components of multiple output media; it delivers 'realistic'
Finally, the simulation results for running multimedia pictures, animation and sound. However, the content of
over the FDDI synchronous channel are discussed. A what is viewed, and the way it is presented is largely
comparison of multimedia over FDDI synchronous and beyond the control of the user. The user interaction is
asynchronous channels is made demonstrating the ability of limited to turning on or off, switching channels, volume
the synchronous channel to provide a near-constant latency control and some image control.
and bandwidth under various loads and configurations. Personal computers demonstrate the other key concept
in multimedia: interaction. Unfortunately, most current
1.1: FDDI Basics personal computers exploit only a limited number of media
such as text, graphics and a few sound tones. It is only now
FDDI is a 100 Mbps high speed local area network that the personal computers' ability to present clearer,
standard developed under the auspices of American sharper images is being exploited by multimedia
National Standards Institute (ANSI) X3T9.5 committee. applications.
Unlike other LANs whose origins were proprietary
products, FDDI was developed by a group of people whose 2.1: Multimedia applications
interest was to create a reliable fault-tolerant, high-speed
network connecting numerous stations over greater The concept of multimedia is only as interesting as the
distances than existing standards. The ANSI X3T9.5 applications it can support. It is the scope of multimedia
committee thus developed a specification for a network applications which is drawing such a tremendous response
based on a dual counter-rotating fiber optic ring using from the information technology industries.
timed-token protocol, which is capable of transmitting data The various multimedia applications are shown in table
at 100 Mbps in each ring and which can extend to 500 1 below.
Table 1: Multimedia applications

APPLICATION MEDIA DESCRIPTION

A=Audio, V=Video,
T=Text, I=Images,
G=Graphics
Education/ Training A, V, T, I A live or taped video instructional session locally at a station or
across a network. This involves transfer of audiovisual and
textual instructions.
Interactive Education/ A, V, T, I, G A live or taped video instructional session locally at a station or
Training across a network. This involves transfer of audiovisual and
textual instructions. The interaction may be via voice, text or
graphics.
Information kiosks A, V, T, G This is dispensing of information such as legal, tourist, consumer
catalogs, dictionaries etc... These kiosks typically have minimal
interaction with the user; mainly instructions and commands.
Banking T, I Number transfers along with documents such as checks, money
orders, internal papers etc...
Medical Info. systems V, I, T, A Stored or real-time capture of X-Rays, CAT scans, reports etc...
with ability to do history, compare and micro and macro
searches.
Library I, T, V Search and retrieval of texts, audio and video cassettes.
Real Estate A, V, I, T Ability to view a neighborhood, block, street, house and interior
of home remotely.
Electronic Mail T, A, I Ability to leave audiovisual messages along with textual
messages.
Home Video Distribution V, A Cable company maintains library of movies which can be
selected and played-back at viewer's convenience.
Travel Agents A, V, I, T Similar in concept to real estate with ability to book travel,
lodging and boarding in one call.
CAD/CAM T, G, A Standard engineering CAD / CAM with the ability to voice-
annotate and have on-line graphics help and notes as pop-up
boxes.
Electronic Collaboration A, V, T Video conferences, concurrent CAD/CAM etc...
Advertising V, A, I, T Ability to reconstruct, edit and create images annotated with
audio and text. The images may be sequenced into a video clip.
Weather I, T Ability to parse satellite images of the atmosphere into weather
reports.

synchronized to the picture. This issue is currently being

addressed in the multimedia community.
There are several issues with multimedia, most notably Most multimedia applications will need network
compression, synchronization and network transparency. support [9]. Videoconferencing [7], distributed training,
Multimedia applications typically consume a lot of etc... require the use of a network. Insofar as the network
disk space for storage and current storage devices such as is concerned, the only difference between multimedia and
Hard Disks, floppy disks, CD-ROMs are not capable of other applications is the integration of voice, video and
storing massive amounts. Moreover, these storage devices data on the same network. Imaging can be modeled as
have a very slow transfer rate (read, write). Motion video bursty data and hence does not require any additional
requiring 75 Mbps transfer rate cannot be sustained using service of the network.
a CD-ROM without some compression techniques. Almost all data transmission is asynchronous in nature.
Synchronization of multimedia objects is an important It is unpredictable and of varying duration. Most local
issue [8]. An object is a unit of data which may be a pixel, area networks are optimized for high throughput, bursty
encoded audio, the multimedia document itself, etc... data transmissions over a shared medium with little or no
Synchronization can be at different granularity's. An latency constraints. If the network is lightly loaded, and a
example of synchronization is voice and video streams; station applies a large load to the network, it will be able
the voice accompanying a video clip needs to be to transmit the load with minimal delays. However, if the
network is heavily loaded, a station applying a large load
will encounter significant delays before transmission.
Thus, although Ethernet under light load offers excellent
average latencies, at high load it offers little or no bound
on the network access time.

Public networks or the telecom networks, on the other

talk-spurt silence
hand, are typically optimized for circuit-switching 650 ms
350 ms
applications such as voice which requires low bandwidth
and low latency. These networks typically cannot provide
low access delay to bursty data. Figure 1: Speech pattern
In order to evaluate the feasibility of using existing
LANs in multimedia applications, we examine the issues
in sound and video transmissions such as bandwidth In digitized voice, the voice signal is sampled at the
requirements, latency, jitter and maximum number of Nyquist rate (twice the signal rate) in order to recover the
sessions. The characteristics of sound and video and the original signal correctly.
requirements of multimedia on networks are examined. Therefore the voice sampling rate is 2*4 khz = 8 khz or
Finally, the feasibility of multimedia over FDDI is one sample every 125 ms in telephony. For stereo sound,
explored. the audio frequencies extend up to 22 khz. Hence
sampling frequencies of up to 44 khz are also used in
3: Definitions digital stereo sound. Each sample is coded into a bit-
stream and the number of code-bits varies depending
Available bandwidth is the bandwidth which is actually upon the coding scheme used.
available for valid transmissions. Available bandwidth
can also be measured in terms of network efficiency. Table 2: Compressed digital audio streams
Thus, in FDDI efficiency h = (T - D) / T; where T= target
token rotation time in ms and D = ring latency in ms. TRANSMISSION RATES bits/sample
If D=0.1 ms and T = 10 ms, then h = 99% and available 2400 bps compressed 1-2
bandwidth is 99 Mbps whereas total bandwidth is 100 64 kbps standard PCM 6-10
Mbps. 40-16 kbps ADPCM scheme[22] 5-2
Latency is the average end to end message delay which 300-400 kbps stereo 8-10
includes time for A/D conversion (if any), sample and
encode, packetization, queuing delay, transmission delay, Although telephone voice is not very bandwidth
propagation delay, receive delay, decode, and intensive, stereo sound requires up to 0.4 Mbps
presentation. bandwidth. At lower bit rates, the audio-quality
Jitter is the maximum instantaneous variation in object deteriorates. Determining the quality of digitized audio is
presentation time. If the object is a packet, then the a very subjective phenomenon and studies have indicated
maximum inter-packet arrival time variation is defined as a wide range of acceptable quality. It also depends on the
packet jitter. application. Spoken voice can be compressed
Session is defined as an interactive communications significantly before it becomes incomprehensible. The
dialogue between two or more users. Thus a telephone study of stereo quality sound is even more subjective.
conversation between two people is a session which There are several coding and compression techniques
consumes a portion of the available bandwidth. available for voice [2], [18], [21], [22]. Adaptive
Differential Pulse Code Modulation (ADPCM) and Digital
4: Characteristics of Sound Speech Interpolation (DSI) are two of the popular
mechanisms for voice compression. DSI is well-suited for
We classified sound as human speech and music. packetized voice transmission as it conserves bandwidth
Human speech or voice is typically in the 0-4 khz during silence intervals in a conversation.
spectrum. The bandwidth of music discernible by the
human ear is 20-25 khz (high fidelity systems have a 4.1: Issues in sound transmissions
bandwidth of 22 khz).
Conversational sound (speech) consists of talk-spurts In voice transmissions [2],[3],[4],[5],[19], bandwidth is
followed by silence periods [3], [5], [6]. The ratio is not an issue in the LAN environment, although it may be
typically 35:65 respectively, with only one person in stereo sound. A critical issue in interactive speech or
speaking at a time. music (or voice mail) played back over the network is
latency. Once a person starts speaking, or music starts to
play, intermittent delays during the speech or music
session, are often noticeable, and sometimes annoying. annoying though coherence is not impaired. For music,
Therefore, once the session starts, it is desirable to the latencies may be more noticeable and hence the delays
maintain a continuous stream of sound. In fact, as shown may be required to be even less in order to be
in the table below, several studies have indicated that imperceptible. Several studies have been conducted in
maximum tolerable latencies for speech are of the order of order to determine effects of network delays on voice
600 ms. Our experiences with satellite communications transmissions [3], [4], [5], [21], [31], [32].
has demonstrated that even 250 ms (one way) delays are

Table 3: Effects of latency on human ear perception

One-way delay Effect of delay

>600 ms Speech becomes incoherent and unintelligible.
600 ms Speech is barely coherent.
250 ms Annoying. Conversation style has to be changed.
100 ms Imperceptible if listener hears from network only and not off the air.
50 ms Imperceptible even if the listener in same room and can hear naturally off the air and from
the network.

Using interactive speech as a model, we decided that network. In an asynchronous network, there is a need to
the maximum end-to-end tolerable latency was 100 ms. buffer. So long as the buffering is not excessive, the jitter
These latencies would be acceptable for a large spectrum has minimal impact on system design. For example, if the
of multimedia applications. maximum packet latency and hence the maximum jitter
The only effect of token jitter is on the need to buffer. between packets were to be restricted to less than 15 ms
In a truly isochronous network (providing constant (as we targeted for the simulation), then
latency), there would be zero buffering requirement in the
the maximum buffering required for a 64 kbps digital
voice stream would be 120 bytes, which is very small.

5: Characteristics of video frames. An NTSC compatible video is 640 x 480 at 30

frames per second. A PAL compatible video is 768 x 516
Video is moving pictures. It is different from imaging at 25 frames per second. The smaller the window size
and graphics mainly in the motion component. Video (fewer number of lines scanned), the lower the bandwidth
represents motion scenes as a rapid sequence of still requirement.

Table 4: Effect of frame rate on human eye perception

Frames per second Effect on human eye

(fps)
<10 fps Eye cannot discern motion. Each frame appears disjointed
12-15 fps Eye can discern motion but is jerky.
30 fps Television quality. Cannot discern high motion component
(blurred); e.g. baseball
60-75 fps HDTV quality. Can discern motion in high-motion games; e.g. ice-
hockey
90 fps Limit of human eye perception
1000 fps Scientific video quality; e.g. shuttle blast-off recording
 The video signal can consume tremendous bandwidth. maintaining similar picture quality. Video bandwidth can
An uncompressed digital NTSC video can consume be reduced in several ways:
anywhere from 90 Mbps to 200 Mbps depending on the
encoding. This is enough to overrun any existing network •Variable resolution
capacity. Compressed video [[28], [29], [30], [32] offers a •Variable frame rate
significant reduction in the bandwidth needs while •Reduced color fidelity
•Removing intra-frame and inter-frame redundancy

Table 5: Video compression techniques, rates and bandwidth requirements

Compression technique frame rate (fps) bandwidth

MPEG [30] 30 1.5 Mbps video stream
MPEG II 60-75 4-10 Mbps
P x 64 [28] 6-15 64 kbps - 2 Mbps

5.1: Issues in video transmissions as those of voice. Depending on the video application,
this delay may vary.
Digital video characteristics have not been studied as We selected 100 ms as an acceptable end-to-end
well as digital voice. It is difficult to characterize latency [31], [32]. We assumed that codecs at each end
compressed video, video codecs and effect of network will consume 30 ms in processing each frame and
delay. The issue of compression algorithms and outputting data (at MPEG rates). This leaves an effective
implementations is beyond the scope of this paper and the 40 ms latency for the network component of the latency.
parameter of most interest to us is the network delay. Of this, 10 ms is the latency across the WAN, and the rest
Using a well-known maxim that " the human eye is is the LAN component of the delay. Assuming
more forgiving than the human ear", we can apply the symmetrical LANs, this leaves 15 ms per LAN acceptable
restriction that video must meet the same delay constraints latency.

WAN
FDDI FDDI

CODEC WAN latency CODEC

latency LAN latency (10 ms.) LAN latency latency
(30 ms) ( 15 ms.) ( 15 ms.) (30 ms)

End-to-end latency (100 ms.)

Figure 2: Latency distribution across the network

If the maximum packet latency and hence the maximum

jitter between packets were to be restricted to less than 15
ms (as we targeted for the simulation), then the maximum
buffering required for a 1.5 Mbps stream would be
approximately 3 Kbytes, which is usual in an FDDI
adapter.
6: Use of FDDI Synchronous class of service 7: Network operation
for video/voice applications
The purpose of our study was to examine the feasibility An FDDI network can be operated in three ways:
of multimedia applications over the FDDI synchronous 1] asynchronous only;
channel when the FDDI is shared with normal, bursty data 2] mixed asynchronous and synchronous;
traffic. The synchronous channel offers a protected, low- 3] synchronous only.
latency bandwidth, which when unused is available to the We decided to test the network in all modes of
normal asynchronous transmissions. operation with a special emphasis on the mixed
A portion of the FDDI bandwidth is allocated to asynchronous and synchronous mode. The offered
synchronous services, either at startup or later by a network load was varied from 80% to about 150% of
bandwidth allocater. Up to 100% of the network capacity. The traffic was a mixture of various applications
bandwidth can be allocated to the synchronous service. In such as voice, video, imaging, file servers, and interactive
other words, it is possible to implement a synchronous data.
only network. This allocation can be fixed, dynamically We wanted to test the network not for its maximum
allocated at session initiation, or on any granularity configurations but for its typical configurations. After
preferred by the network administrator. A standard conducting a survey of the existing implementations and
allocator such as CCITT Q.931, may be used to perform practices, we concluded that a maximum unsegmented
call-setup, tear-down and bandwidth allocation and network was in the range of 40-60 nodes. An
monitoring. SMT defines a protocol and several MIB unsegmented network has nodes on the same physical
attributes which can be used to monitor and control the cable-plant with no intervening bridges, routers or some
bandwidth allocation [27]. such interworking units. Usually, networks do not exceed
Each multimedia station is allocated a portion of the 50 nodes because of issues such as loading, administrative
synchronous bandwidth. In order to allocate the domains, and traffic isolation. We selected a network
bandwidth, each station needs to characterize the with nodes in the range of 48 to 55 as the representative
application requirements in terms of overhead and maximum of the typical network.
payload, where overhead includes token capture, framing
and higher layer protocol headers, and payload is the 7.1 Topology
actual synchronous data (e.g. voice, video). This should
be calculated in units of bytes per 125 microseconds. An The following topology was adopted as the model for
application of 1.5 Mbps would require 1.5 x 106 x 125 x the study. A LAN-WAN-LAN model was seen as
appropriate for typical multimedia services. We assumed
10-6 /8= 23.4375 rounded up to 24 units of bandwidth. A that the two LANs were symmetrical in behavior.
similar calculation should be done for the overhead. The
total bandwidth available is of the order of 100 x 106 x
125 x 10-6 /8 = 1562 units. Following the bandwidth
allocation, it is necessary to select the packet sizes for the
negotiated TTRT. For example, with a TTRT of 8 ms, the
packet size for the above synchronous traffic (1.5 Mbps
stream) can be calculated as the number of bytes that the WAN
stream will produce in 8 ms. This is 1500 bytes. Hence, FDDI FDDI
with an 8 ms TTRT, and a 1.5 Mbps video stream, 1500
byte packets will be transmitted per token rotation.
The above method of allocating bandwidth was
selected so that an application would not have to change
bandwidth allocation every time that the TTRT value concentrator/
changed. Only the packet size would change while station
maintaining a constant overhead. There are other Figure 3: Topology for simulation
mechanisms for allocating bandwidth which may be
simpler and more suitable for different network
configurations. A single hierarchy of connections was selected because
For the purposes of the simulation we allocated of the linearity of each LAN segment. Therefore, the
synchronous bandwidth based on the TTRT value. Thus delay characteristics for a single LAN could be linearly
an 8 ms TTRT with a 1.5 Mbps application would require scaled to represent multiple-level LAN hierarchy. The
1500 byte transmission time or 1500 x 80 ns = 0.12 ms per WAN connection was assumed to offer a fixed latency
TTRT. path. The WAN could consist of ISDN, fractional T1, T1
or T3 lines depending on the bandwidth required.

7.2: Ring latency

 A ring latency of 84 ms was used to test a small ring. 7.4: Traffic and service models
This corresponds to roughly 7 kilometers of cable. To test
the behavior of the network with larger ring latencies, a Of the 50 stations, 26 were set-up to be voice/video
ring latency of 1 ms was selected. This corresponds to stations, 3 were low-rate imaging sources, 10 were
roughly 190 kilometers of cable. interactive data terminals, and 10 were file-servers.
Optionally, high burst-rate imaging sources (1 to 7
7.3: Target Token Rotation Time stations) were used in place of the 3 low-rate imaging
sources. The traffic was thus split into voice/video,
We decided to operate with three values of TTRT- 8 imaging and data. Imaging and data were further sub-
ms, 16 ms, and 24 ms TTRT values less than 8 ms were divided. There were two models for the imaging and two
not selected because network efficiency drops models for the data.
significantly [33]. TTRT values larger than 24 ms were The following table shows the characteristics of the
not selected because the packet latencies would be various traffic models that we selected.
unacceptable for the multimedia traffic. Additionally, for
a synchronous only network, a TTRT of 26 ms was used.

Table 6: Traffic distributions

TRAFFIC TYPE INTER- PACKET Peak OFFERED Avg. OFFERED BUFFER

ARRIVAL LENGTH LOAD LOAD SIZE
TIME (in ms) (in bytes)1 (in Mbps)2 (in Mbps) (in packets)3
Imaging host 0.33 4K + 256 106.25 10 1000
Imaging workstation 3.6 4500 10 2 10
file data 37 1500 + 256 10 3 50
interactive terminal data 40 500 - 0.1 10
voice4 16, 84.5 64, 2028 + 20 + 256 0.032, 0.218 0.0112, 0.076 10, 10
video 8,14,16,24 1500, 2304, 3000, 4500 1.5, 1.316, 1.5, 1.5 1.5, 1.316, 1.5, 1.5 10
gateway5 - - 10 10 50

Overall, we stressed the network with a variety of traffic 7.4.2: Voice streams
models to ensure that the network is robust under extreme
operating conditions. Associated with each video stream is a packet voice
stream. This stream was characterized as interactive and
7.4.1: Video streams stored (MPEG CD-ROM specifications).
The interactive voice is characterized as a sequence of
Each video source represents a compressed video 32 kbps voice packets, each 64 bytes in size. The voice
stream. An MPEG-type stream rate of 1.5 Mbps is used. source is modeled as a sequence of on-off periods,
This consists of a train of packets ranging from 4500 bytes representing the talk-spurts and silence periods typical in a
to 1500 bytes. The overall video loading is characterized voice conversation. The ratio of talk to silence periods is
through a parameter which represents the number of 35: 65. The resulting offered load of the voice stream is
simultaneously active video sources. 11.2 kbps.
A prescribed fraction of the video stream is directed The stored voice is characterized as a sequence of 2
through a bridge or a gateway to another FDDI or WAN Kbyte packets (2028 + 20 MPEG headers), with an inter
network. arrival rate of 84.5 ms. These packets are typically
The target video latency for a video packet across a interleaved with the video stream in a ratio of one voice
single FDDI is set to be around 15 ms (99% of packets) every six video packets. On the FDDI, these packets are
and of the order of 10 ms in the average. repacketized if necessary to smaller packet sizes.

1 The length denotes the data + headers. The headers were deliberately chosen to be a large number.
2 The arrival rate for some traffic had a distribution model rather than a constant rate, which led to peak offered loads and average offered
loads.
3 The buffer size corresponds to the buffering at the transmit and receive queues. If an incoming packet finds the buffer full, it is dropped.
This corresponds to blocking.
4 An interactive voice model (32 Kbps ADPCM), and a MPEG stored voice-stream model were selected for modelling.
5 The gateway loading was approximately equal to 4 voice/video stations and 1.3 fileservers.
 The desired latency for voice packets across a single In an asynchronous only network, with no synchronous
FDDI is also set to be approximately 15 ms (99% of bandwidth allocated or used, the voice and video are
packets) and of the order of 10 ms in the average. treated as data. No separate queue is allocated on transmit
or receive. In such a network also it is possible to
7.4.3: Data : Interactive maintain a bound on the delay suffered by the packets.
The following observations refer to figures 4 to 19.
This traffic consists of short packets (500 bytes)
arriving at random with low response time requirements of 8.1.1: Effect on 99% latencies
50 ms (99% of packets) and 25 ms in the average.
Due to the unpredictable nature of the traffic
7.4.4: Data -File transfer (asynchronous and bursty), the delay cannot be tightly
bounded. As can be seen from the figures 8 and 10, the
This is modeled as file data to/from Ethernet hosts with 99% latencies suffered by video packets is as high as 48
the FDDI being used as a backbone to Ethernet clients. ms when the network is not overloaded but running close
The file length is assumed to be uniformly distributed to capacity (90% load). When the network is overloaded,
between 1500 and 25000 bytes. Each file on average the latencies can be as high as 252 ms.
consists of 8 packets of 1500 bytes + 256 bytes headers, In a more typical environment, where the traffic does
each arriving at 10 Mbps once the file transfer is initiated. not consist of such high burst sources (imaging at 100
The file inter-arrival time is exponential. The offered Mbps), it is possible to obtain low latencies. We were
load is 3.6 Mbps. able to verify this in our simulation (see figure 6) where in
an asynchronous only network, with 86% network
7.4.5: Imaging : low-end (workstation) loading, and a TTRT of 24 ms voice and video packet
latencies were restricted to under 15 ms.
This traffic source was used in some of the simulation
runs. 8.1.2: Effect of TTRT on latencies
This is modeled as images coming off Ethernet hosts
into the FDDI host at 10 Mbps. The image size In an overloaded network the higher the TTRT value,
distribution is uniform over 1.25 - 5 Mbytes. This image lower the latencies. In the 8 to 24 ms range, it was
stream is packetized into maximum length FDDI packets observed that the 24 ms TTRT value consistently offered
(4096 + 256 bytes). The image inter-arrival time is varied lower delays for all traffic types when the ring was large.
and a default value of 20% on-time and 80% off-time is (figure 8 and 10).
assumed. This assumes that a host is busy with imaging
only 20% of the time. Thus the peak offered load is 10 8.1.3: Effect of ring size
Mbps, but the average offered load is 2 Mbps.
The maximum acceptable delay in transmitting an Increasing ring latencies had an adverse effect on the
image and receiving it at the receiver is 1 s (99% of packet latencies. This was reflected in the increase in the
packets) and 0.5 s average delay. A buffer size of 10 mean and maximum latencies for voice and video (figure
packets is used. Any over-flow leads to packet dropping. 6 and 10). This effect was less noticeable in the 99%
latencies.
7.4.6: Imaging : High-end (host)
8.1.4: Effect of buffer sizes
This traffic source is used to simulate the impact of a
very bursty load on the network. The image distribution is Buffer sizes were allocated to the individual queues,
uniform over 1.25 to 5.625 Mbytes. A single image asynchronous and synchronous at different stations.
stream consists of regularly arriving maximum sized Therefore the asynchronous buffer size at the imaging
packets (4096 + 256 bytes). The average image inter station was varied from 10 to 1000 packets, at the file
arrival rate is 0.32768 ms. The peak offered load is server stations it was 50 packets, and at the interactive
106.25 Mbps and the average offered load is 10 Mbps. terminals it was 10 packets. Every synchronous station
had a 10 packet buffer.
8: Results These buffers never overflowed except in the overload
scenario. Then too, the blocking or buffer overflow was
The following sections summarize the results of the occurring only at the imaging station. This result is
simulations. intuitive because the imaging stations were offering
instantaneous overload. This burst would fill up the
8.1: Case 1- Asynchronous only network buffers and since the network was not faster than the
application, the buffers would not empty out fast enough.
At heavy loads these buffers are unable to empty and
hence lead to overflow. Low burst-rate applications such
as file service and interactive terminal traffic were not latencies are within 24 ms. For 8 ms and 16 ms TTRT
capable of exceeding the network capacity and hence the values, the latencies are always within 16 ms. Even under
network could always clear the buffers faster than the extreme stress, the synchronous channel offered a low-
application could fill them. latency path for time-critical applications such as
The percentage of traffic blocked due to overflow multimedia.
decreased slightly with increasing TTRT. This is again
intuitive because the larger TTRT values allowed longer 8.2.2: Effect of TTRT on latencies
transmission times which in turn allowed the buffers to be
drained more often. Within the range of TTRT values yielding acceptable
latencies, it was more difficult to isolate the better TTRT.
8.1.5: Effect on gateway We observed that while 8 ms TTRT yields excellent
values for voice/video traffic, the asynchronous bursty
For traffic from FDDI to other LANs or WANs, the traffic suffered lower latencies at the higher TTRT values.
latency was gated by the characteristics of the other LAN Considering all traffic streams, we observed that a 16 ms
or WAN. So long as the outbound traffic was less than the TTRT offered better all-round latencies in a mixed
capacity of the WAN or LAN, there was little or no synchronous and asynchronous network (figure 5, 7, 9 and
queuing delay. The major component of the delay was 11).
then the transmission delay.
The effect of the changing of the various network 8.2.3: Effect of ring sizes
parameters on the gateway (figures 4-18) was similar to
that of other end-stations except that in the overload case To observe the effect of ring sizes on the network
the gateway suffered significant blocking. The gateway performance, we simulated with a small ring size of 84 ms
was the bottleneck for voice/video sessions spanning the latency and a large ring size of 1000 us latency. The
LAN-WAN-LAN connection. The minimum latencies effect of increased ring sizes on voice / video (VV) is
observed when the network loading was 90% and with at readily apparent on the mean delays. The mean latencies
least one high burst-rate source on the network, was 24 ms increased by as much as 40% whereas the 99% latencies
at 8 ms TTRT. increased by about 10-15% only. Since the important
When the high burst-rate source (imaging with peak parameter for system design is the 99% latency rather than
offered load of 100 Mbps and average of 10 Mbps) was the mean latency, this implies that in the range of typical
removed, the gateway provided acceptable performances ring sizes (50 ms to 400 ms ), the 99% latencies are fairly
with latencies less than 15 ms. constant.

8.2: Case 2- Asynchronous plus synchronous 8.2.4: Effect of buffer sizes

network
The effect of buffer sizes is slightly higher blocking
In this network the voice and video are transmitted than in the asynchronous only network. This is due to the
over the synchronous channel. The synchronous synchronous traffic effectively blocking some portion of
bandwidth is allocated per station based on the application the bandwidth. If 20% of the bandwidth is allocated to the
requirement. The voice and video packet sizes are varied synchronous channel, then for the asynchronous
according to the TTRT requirements. If 64 byte applications FDDI appears to be a 80 Mbps asynchronous
interactive voice packets are used, then the packet size is data pipe. This is an excellent result because it implies
constant for the different TTRT. For video, the packet- that if the asynchronous application capacity requirement
size is 1500 bytes for 8 ms, 3000 bytes for 16 ms and 4500 is known, the rest of the FDDI bandwidth can be allocated
bytes for 24 ms TTRT. The gateway is used for to the synchronous channel and there will be little or no
voice/video and file server data forwarding. The gateway effect on the asynchronous applications.
is allocated synchronous bandwidth in proportion to the
traffic leaving the LAN. If four video streams are leaving 8.2.5: Effect on gateway
the LAN, then the gateway is allocated bandwidth equal to
four video streams (e.g. at 0.75 ms per video stream the For traffic from FDDI to other LANs or WANs, the
gateway is allocated 3 ms). latency was gated by the characteristics of the other LAN
or WAN. The major component of the delay was then the
8.2.1: Effect on 99% latencies transmission delay.
The effect of the changing of the various network
We observe that the 99% latencies for voice/video parameters was similar to that of other end-stations
streams is fairly constant and under all circumstances- (figures 5, 7, 9,11, 13, 15, 17). It was observed that for
90% and 150% loading on a small ring, 90% and 150% the incoming traffic from the gateway, the latencies were
loading on a large ring, and different TTRT - the 99% less than 16 ms for TTRT values of 8 and 16 ms, under
overload. For 24 ms TTRT the delay was of the order of
23 ms under overload. In light to heavy loading video, imaging and data communications services can be
conditions the latency was 8 ms. integrated over FDDI without compromising any service
requirement in all but the most extreme cases. These
We also observed that changing the number of results demonstrate the feasibility of FDDI as an
outbound sessions had little or no impact on the latency if integrated network and are part of a continuing study on
the appropriate synchronous bandwidth was allocated. multimedia over FDDI.

8.2.6: Synchronous bandwidth allocation 10: Acknowledgments

It was observed that incorrect bandwidth allocation Our acknowledgments to Bob Grow (XDI), J. D. Russell
could dramatically affect the performance. Allocating the (IBM), Bob O'Hara, Basil Alwan and Dave Roberts
appropriate bandwidth resulted in a guaranteed low- (Advanced Micro Devices) for their invaluable help and
latency channel for the application. suggestions in the model development.

8.3: Synchronous only network

We simulated a synchronous only network with 70
voice/video stations and a ring latency of 122 ms (figure
19). In this configuration, the only traffic on the network
was voice/video. Each station was generating voice and
video traffic at a combined rate of 1.5 Mbps. Each station
was allocated 0.37 ms per token rotation leading to a
TTRT of 26 ms. The total load was near the maximum
sustainable by the network.
The results showed that the 99% latencies were less
than 7 milliseconds. In fact, even though the TTRT was
26 ms ( in order to accommodate 70 stations), the actual
maximum delay suffered by any packet was of the order
of 10 milliseconds.
It was also observed that there was a sharp drop in
performance with a small increase in the number of
stations. This is because the network is operating at
maximum capacity and even a slight increase in load can
cause the synchronous bandwidth to be over allocated.
We concluded that it is possible to calculate the maximum
number of synchronous stations that can be supported and
provide acceptable latencies. For the given application
(1.5 Mbps MPEG stream), it is between 55-65 stations,
depending on the safety factor desired.
It is not advisable to apply bursty and asynchronous
traffic to the synchronous channel as it can lead to
extremely high delays [34], [35]. This is because a burst
offers instantaneous load which can exceed the allocated
bandwidth causing the queuing component of the delay to
increase significantly .

9: Conclusion
The FDDI asynchronous mode provides excellent
multimedia services at normal loads. However, it has
limitations under heavy and extremely bursty loads which
may not be acceptable for voice/video services. The FDDI
synchronous mode of transmission provides a near
constant, low latency service under various loads and
configurations. Our results demonstrate that a large
number of simultaneous MPEG and Px64 sessions can be
supported even when the network is in over-load . Audio,
11: Bibliography [18] N. S. Jayant & S. W. Christensen, "Effects of Packet
Losses in Waveform Coded Speech," IEEE Trans. on
[1] I. Brackenbury & H. Sachar, "The road to Communications, COM-299(2), Feb '81
multimedia," Future Generation Computer Systems 7, [19] M. Frontini & G. Watson, "An investigation of
1991, pp. 91-96 packetised voice on the FDDI token ring", HP technical
[2] K. Sriram, R. S. McKinney & M. H. Sherif, "Voice report, no. HPL-BRC-TR-87-019
Packetization and Compression in Broadband ATM [20] V. S. Frost, W. W. LaRue, et. al, "A tool for local area
Networks," IEEE JSAC, vol. 9, no. 3, Apr. 91, pp. 294- network modeling and analysis," SIMULATION, nov. '89,
303 pp. 283-297.
[3] T. Bially, A. J. McLaughlin, & C. J. Weinstein, "Voice [21] D. O. Bowker, C. B. Armitage, "Performance issues
communication in integrated digital voice and data for packetized voice communications," Proc. Nat.
networks," IEEE Trans. Commun., vol. COM-28, Sep. 80, Commun. Forum, vol. 41, n0. 3, '87, pp. 1087-1092
pp. 1478-1490 [22] M. H. Sherif, D. O. Bowker, et. al., "Overview of
[4] J. G. Gruber, "Delay related issues in integrated voice CCITT embedded ADPCM Algorithms," Proc. IEEE
and data networks," IEEE Trans. Commun., June 81, pp. Supercomm/ICC'90, Atlanta, GA, Apr. '90, vol. 3, pp.
787-800 1014-1018
[5] J. G. Gruber, "Performance Requirements for [23] T. D. C. Little & A. Ghafoor, "Network
integrated Voice/Data networks," IEEE JSAC, DEC. 83, Considerations for Distributed Multimedia Objects
pp. 981-1005 Composition and Communication," IEEE Network
[6] P. T. Brady, "A statistical analysis of on-off patterns in Magazine, Nov. '90, pp. 32-49
16 conversations," Bell Syst. Tech. J., vol. 47, Jan. '68 [24] FDDI Token Ring Media Access Control," American
[7] K. Watabe, S. Sakata,, et. al, "Distributed Desktop National Standard for Information Systems, X3.139-1987
Conferencing System with Multiuser Multimedia [25] FDDI Physical Layer Protocol, American National
Interface," IEEE JSAC, vol. 9, no. 4, may 91, pp. 531-539 Standard for Information Systems, X3.148-1988
[8] R. Steinmetz, "Synchronization Properties in [26] FDDI Physical Layer, Medium Dependent (PMD),
Multimedia Systems," IEEE JSAC, vol. 8, no. 3, Apr. 90, American National Standard for Information Systems,
pp. 401-412 X3.166-1990
[9] C. Nicolaou, "An Architecture for Real-Time [27] FDDI Station Management (SMT), Preliminary Draft
Multimedia Communication Systems," IEEE JSAC, vol. proposed American National Standard, Rev. 7.1,
8, no. 3, Apr. 90, pp. 391-400 X3T9/92-037
[10] D. Dykeman & W. Bux, "Analysis and Tuning of the [28] CCITT Recommendation H.261, Line Transmission
FDDI Media Access Control Protocol," IEEE JSAC, vol. on Non-telephone Signals: Video Codec for Audiovisual
6, no. 6, July 88 services at p x 64 kbps
[11] R. M. Grow, "A timed token rotation protocol for [29] ISO/IEC DIS 10918-1; Information technology -
local area networks," Electro '82, Token Access Protocols, Digital compression and coding of continuous-tone still
May 82 images; a.k.a JPEG
[12] M. J. Johnson, "Reliability mechanisms of the FDDI [30] ISO/IEC JTC 1/SC 29 N 071; Coded representation of
high bandwidth token ring protocol," Comput. Networks picture, audio and multimedia/hypermedia information;
ISDN Syst., voi. 11, Feb. 86, pp. 121-131 a.k.a MPEG
[13] M. J. Johnson, "Proof that timing requirements of the [31] J. D. Russell, "Trends in Communication Throughput
FDDI token ring protocol are satisfied," IEEE Trans. Demand," Fibertour/ComputerNet Conf., Boston, 10/90
Commun., vol. COM-35, June 87, pp. 620-625 [32] J. D. Russell, "Communications Performance
[14] F. E. Ross, "FDDI - a Tutorial," IEEE Requirements in Multimedia Applications"
Communications, May '86, 24(5) [33] R. Jain, "Performance Analysis of FDDI Token Ring
[15] V. Iyer & S. Joshi, "FDDI's 100 Mbps Protocol Networks: Effect of Parameters and Guidelines for setting
Improves on 802.5 Spec's 4 Mbps LImit," EDN, May 2 TTRT," IEEE Lightwave Telecommunications Systems,
'85, pp. 151-160 May '91, pp. 16-22
[16] R. K. Moulton, "FDDI - a 100 Mps Solution on [34] W. Genter & K. S. Vastola, "Delay Analysis of the
Fiber," Lightwave, Oct '86, pp. 47-54 FDDI Synchronous Data Class", IEEE
[17] H. R. Miller, P. Zafiropulo & F. Closs, "Data/Voice [35] P. Martini & T. Meuser, "Service Integration in
integration based on IEEE 802.5 Token-Ring LAN, FDDI," Proceedings of the 15th Conference on Local
EFOC/LAN 86 Proceedings, June '86, pp. 66-76 Computer Networks, 1990, IEEE Computer Society Press
 Mean,99% Delay (20,21,22- Image WS Not Shown)

Figure 4: Async. only n/w, normal load, 84 microsec. ring TTRT (ms)

8
Mean
16
OfLd .9, Async, 16.8km. 24
8
99%
16
24
70

60

50

D
e 40
l
a
y

m 30
s

20

10

0
Voice Video Interactive File Data Video Gtw
 Mean,99% Delay (23,24,25)

Figure 5: Sync./Async. n/w, normal load, 84 microsec. ring with imaging ws TTRT (ms)

8
Mean
16
OfLd .9, Async/Sync, 16.8km. 24
8
99%
16
24
10000
TTRT 8- SBA: .25(VV), 1.0(GW

TTRT 16- SBA: .5(VV), 2.0(GW

TTRT 24- SBA: .75(VV), 3.0(GW

1000

D
e
l
a
100
y

m
s

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Mean,99% Delay (20,21,22)

Figure 6: Async. only n/w, normal load, 84 microsec. ring with Imaging TTRT (ms)
WS

Mean 8
16
24
8
99%
OfLd .9, Async, 16.8km. 16
24
10000

1000

D
e
l
a
100
y

m
s

10

1
Voice Video Imaging Interactive File Data Video Gtw
WS
 Figure 7: Sync./Async. n/w, normal load, 84 microsec. ring TTRT (ms)

8
Mean
16
24
8
OfLd .9, Async/Sync., 16.8km. 99%
16
24
70

60 TTRT 8- SBA: .25(VV), 1.0(GW

TTRT 16- SBA: .5(VV), 2.0(GW

TTRT 24- SBA: .75(VV), 3.0(GW

50

D 40
e
l
a
y

m
s 30

20

10

0
Voice Video Interactive File Data Video Gtw
 Figure 8: Async only n/w, normal load, 1000 microsec. ring TTRT (ms)

Mean 8
16
24
OfLd .9, Asyn, 200km. 8
99%
16
24
130

120

110

100

90

80
D
e
l 70
a
y
60
m
s
50

40

30

20

10

0
Voice Video Interactive File Data Video Gtw
 Mean, 99% Delay (35,36,37- Image WS not shown)

Figure 9: Sync/Async n/w, normal load, 1000 microsec. ring

120

110 OfLd .9, Async/Sync, 200km

TTRT 8- SBA: .22(VV), .88(GW)

TTRT 16- SBA: .488(VV), 1.95(GW)

100
TTRT 24- SBA: .75(VV), 3.0(GW)

90

TTRT (ms)

80

70 Mean
D
e 8
l
16
a
60 24
y
8
m 16
s 99%
50 24

40

30

20

10

0
Voice Video Interactive File Data Video Gtw
 Figure 10: Async only n/w, normal load, 1000 microsec. ring with Imaging WS TTRT (ms)

Mean 8
16
24
OfLd .9, Asyn, 200km. 99% 8
16
24
10000

3584

2048

1000

D
e 128
l 86.4
a
100
y
51.2
64
m
s
38.4
32

10

1
Voice Video Imaging Interactive File Data Video Gtw
WS
 Figure 11: Sync + Async n/w, normal load, 1000 microsec. ring with Imaging WS TTRT (ms)

Mean 8
16
OfLd .9, Async/Sync, 200km. 24
99% 8
16
24
10000

3840 TTRT 8- SBA: .22(VV), .88(GW)

TTRT 16- SBA: .488(VV), 1.95(GW)

2150

1792 TTRT 24- SBA: .75(VV), 3.0(GW)

1000

D
e
l 115
a
100
y
51
m
s

22.4
22.4 22.4

14.4 14.4

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Figure 12: Async. only n/w, overload, 84 microsec. ring
TTRT (ms)

8
Mean
16
OfLd 1.5, Async, 16.8km. 24
8
99%
16
24
370
360
350
340
330
320
310
300
290
280
270
260
250
240
230
220
D
210
e
l 200
60%bl
a 190
55%bl
y 180
170 60%bl
m 4%bl
160
s 13%bl
150 16%bl
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Voice Video Interactive File Data Video Gtw
 Figure 13: Sync + Async n/w, overload, 84 microsec. ring
TTRT ms

8
Mean
16
OfLd 1.5, Async/Sync, 16.8km. 24
8
99%
16
24
320
310
300 TTRT 8- SBA: .25(VV), 1.0(GW
290
280 TTRT 16- SBA: .5(VV), 2.0(GW
270
260 TTRT 24- SBA: .75(VV), 3.0(GW

250
240
230
220
210
200
190
D
e 180
l 170
a
160
y 8%bl
150
7%bl
m 140
s 15%bl
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Voice Video Interactive File Data Video Gtw
 Figure 14: Async. only n/w, overload, 84 microsec. ring with Imaging WS
TTRT (ms)

8
Mean
16
OfLd 1.5, Async, 16.8km.
24
8
99%
16
60%bl 24
50%bl
10000 45%bl

1000

4%bl
13%bl
16%bl

D
e
l
a 60%bl
100
y 55%bl

m 60%bl
s

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Figure 15: Sync. + Async n/w, overload, 84 microsec. ring
TTRT ms

8
Mean
16
OfLd 1.5, Async/Sync, 16.8km. 24
8
99%
16
24
100000

TTRT 8- SBA: .25(VV), 1.0(GW

55%bl
60%bl TTRT 16- SBA: .5(VV), 2.0(GW
55%bl
TTRT 24- SBA: .75(VV), 3.0(G
10000

1000
D
e 8%bl
l 7%bl
a
15%bl
y

m
s
100

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Figure 16: Async only n/w, overload, 1000 microsec. ring with Imaging WS
TTRT (ms)

8
Mean
16
OfLd 1.5, Async, 200km. 24
8
99%
16
24
100000

65%bl

10000 61%bl
50%bl

1000
D
e 14%bl
l 12%bl
a 66%bl
19%bl
y 62%bl
62%bl
m
s
100

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Figure 17: Sync + Async n/w, overload, 1000 microsec. ring with Imaging WS
TTRT (ms)

8
Mean
16
OfLd 1.5, Async/Sync, 200km. 24
8
99%
16
24
100000

65%bl
64%bl
58%bl

10000

1000 25%bl
D 12%bl
e 18%bl
l
a
y

m
s
100

10

1
Voice Video Imaging WS Interactive File Data Video Gtw
 Figure 18: Async only n/w, normal load, 1000 microsec. ring
TTRT (ms)

8
Mean
16
24
OfLd 1.5, Async, 200km.
8
99%
16
24
430
420
410
400
390
380
370
360
350
340
330
320
310
300
290
280
270
260
D 250
e 240
230 66%bl
l
a 220 62%bl
y 210 14%bl 62%bl
200 12%bl
m 190 19%bl
s 180
170
160
150
140
130
120
110
100
90
80
70
60
50
40
30
20
10
0
Voice Video Interactive File Data Video Gtw
 Figure 19: Sync only n/w, normal load, 122 microsec ring

Mean
OffLd .96, Sync Tfc, 24km. 99%
All VV, All Synch
7

70 Stations Suppo

Utilizing 96.8 Mbps

6
TTRT: 26 ms

SBA: 0.37 ms

D4
e
l
a
y

m
s 3

0
Voice Video