1, MARCH 2003
Abstract—We design an interactive video-on-demand (VOD) sign can provide an ideal interactive service to the customers,
system using both the client–server paradigm and the broad- but it needs dedicated resources (such as I/O bandwidth) to
cast delivery paradigm. Between the VOD warehouse and the maintain a video stream for each ongoing customer. For large
customers, we adopt a client–server paradigm to provide an inter-
active service. Within the VOD warehouse, we adopt a broadcast
scale applications with many customers, this design requires
delivery paradigm to support many concurrent customers. In large amount of resources.
particular, we exploit the enormous bandwidth of optical fibers A client–server design can use a batching policy [4]–[8] to
for broadcast delivery, so that the system can provide many video serve more concurrent customers. The main idea is that the
program and maintain a small access delay. In addition, we design system waits for a time interval (called batch window) to collect
and adopt an interleaved broadcast delivery scheme, so that every a batch of requests for a video program. Then the system creates
video stream only requires a small buffer size for temporary
storage. A simple proxy is allocated to each ongoing customer, and one video stream for this program and multicasts it to a batch of
it retrieves video from the optical channels and delivers the video customers. In this manner, one video stream can serve multiple
to the customer through an information network. The proposed customers simultaneously. However, the customers have to wait
VOD system is suitable for large scale applications with many before starting a VOD session (the waiting time is called access
customers, and it has several desirable features: 1) it can be scaled delay) and they cannot perform (or can only perform some con-
up to serve more concurrent customers and provide more video
programs, 2) it provides interactive operations, 3) it only requires
strained) interactive operations. Several batching policies have
point-to-point communication between the VOD warehouse and been proposed in the literature and they are as follows.
the customer and it does not involve any network control, 4) it has • Dan et al. [5] proposed that when the system can establish
a small access delay, and 5) it requires a small buffer size for each a new video stream, it selects the batch with the largest
video stream. number of waiting customers and creates a video stream to
Index Terms—Broadcast delivery paradigm, client-server pro- serve all the customers in this batch. This batching policy
gram, video-on-demand. can minimize the mean access delay, but some customers
may experience long access delay.
I. INTRODUCTION • Dan et al. [6] proposed to choose a shorter batch window
for the more popular video programs. They developed an
(a)
(a)
(b)
Fig. 4. Basic broadcast delivery for VOD. (a) Proxy retrieves the shaded pages. (b) Proxy delivers the retrieved pages to the customer.
Therefore, it must have buffer for temporary storage. To deter- livery rate, the buffer occupancy is increasing with
mine the buffer size required, we analyze the buffer occupancy time until it reaches a maximum at the end of this
in the proxy. Fig. 4(c) shows the buffer occupancy versus time, slot. The maximum buffer occupancy is equal to
as follows. retrieval rate delivery rate duration of a slot
• In the first slot of the first cycle, the proxy retrieves .
page 1 and delivers it to the customer simultane- • In the subsequent slots, the proxy does not re-
ously. Since the retrieval rate is faster than the de- trieve any page but it still delivers the remaining portion
134 IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 5, NO. 1, MARCH 2003
(c)
Fig. 4. (Continued.) Basic broadcast delivery for VOD. (c) Buffer occupancy in the proxy versus time. The buffer size required by the proxy is R T.
(a)
Fig. 6. Interleaved broadcast delivery for VOD. (a) Proxy retrieves the shaded pages.
channel, and their minipages are interleaved over nine minis- , , , , in a similar manner, and the details are shown
lots. Specifically, we broadcast the minipages of pages 1, 2, in Fig. 6(a).
and 3 in channel 1, as follows: After retrieving all the minipages from channel 1, the proxy
• broadcast their first minipages one after the other (i.e., tunes its receiver to channel 2. It retrieves minipage in the
minipage , then , and then ); fourth minislot of the fourth cycle, then retrieves minipage
• broadcast their second minipages one after the other (i.e., in the seventh minislot of the fourth cycle, and then retrieves
minipage , then , and then ; minipage in the first minislot of the fifth cycle. The proxy
• broadcast their third minipages one after the other (i.e., retrieves minipages , , , , , in a similar fashion,
minipage , then , and then ). and the details are shown in Fig. 6(a).
In channel 2, we broadcast the minipages of pages 4, 5, and 6 After retrieving all the minipages from channel 2, the
in a similar fashion but in a different order: broadcast their third proxy tunes its receiver to channel 3 to retrieve minipages
minipages, then their first minipages, and then their second . The details are shown in Fig. 6(a).
minipages. In channel 3, we broadcast the minipages of pages We note that the proxy waits for minislots after retrieving
7, 8, and 9 in another order: broadcast their second minipages, the last minipage of a page, and it waits for minislots
then their third minipages, and then their first minipages. after retrieving any other minipage. To ensure continuous video
Fig. 6(a) shows how the proxy retrieves the minipages from delivery, a page (or minipages) must last for one cycle and
the optical channels. In the first cycle, it retrieves minipage one minislot.
in the first minislot, then retrieves minipage in the fourth The proxy delivers the minipages to the customer as follows.
minislot, and then retrieves minipage in the seventh minislot. When it starts to retrieve a minipage, it starts to deliver this
In the second and third cycles, the proxy retrieves minipages , minipage to the customer. It continuously delivers the minipages
136 IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 5, NO. 1, MARCH 2003
TABLE I
BASIC BROADCAST DELIVERY VERSUS INTERLEAVED BROADCAST DELIVERY
(a) (b)
(a)
Fig. 7. Pause operation. (a) Ideal pause operation. (b) Approximate pause
operation.
(a) (b)
Fig. 8. Fast forward operation. (a) Ideal fast forward operation.
(b) Approximate fast forward operation. (b)
Fig. 9. Realization of approximate fast forward operation; the proxy only
much smaller buffer size per proxy. In addition, as we will retrieves and delivers the shaded pages. (a) Realization at the page level.
(b) Realization at the minipage level.
explain in Section IV, it can also support better interactive
operations.
played at a faster and constant rate. Fig. 8(b) shows the approx-
IV. PROVISION OF INTERACTIVE OPERATIONS imate fast forward operation. It plays a small portion of video at
the normal rate, then skips a portion, then plays a small portion,
In this section, we describe a set of interactive operations and then skips a portion, etc.
which are approximations of the ideal ones, and explain how The approximate fast forward operation can be realized at the
the proposed VOD system can provide these approximate oper- page level or minipage level. If it is realized at the page level,
ations. Our explanation will be based on the interleaved broad- it plays a page and then skips some subsequent pages [e.g.,
cast delivery scheme, but the ideas are also applicable to the see Fig. 9(a)]; if it is realized at the minipage level, it plays a
basic broadcast delivery scheme. minipage and then skips some subsequent minipages [e.g., see
Fig. 9(b)]. In general, if the cycle duration is shorter, the approx-
A. Pause imation is closer to the ideal one. In particular, the realization at
Fig. 7(a) shows the ideal pause operation. When the customer the minipage level is better than that at the page level because a
issues a pause command, the playout point of the video program minipage lasts for a shorter duration.
is “frozen,” and the pause duration can be any positive value.
Fig. 7(b) shows the approximate pause operation, in which the C. Fast Rewind
pause duration is an integral multiple of the cycle duration . If Fig. 10(a) shows the ideal fast rewind operation. When the
is smaller, the approximate pause operation is more similar customer issues a fast rewind command, the video program is
to the ideal one. played in the reverse order at a fast and constant rate. Fig. 10(b)
The proposed VOD system can provide the approximate shows the approximate fast rewind operation. It plays a small
pause operation as follows. When a customer issues a pause portion of video at the normal rate, then skips a previous portion,
command to the proxy at time , the proxy stops retrieval/de- then plays a small portion, and then skips a previous portion, etc.
livery temporarily. When the customer issues a resume The approximate fast rewind operation can be realized at the
command at time , the proxy resumes retrieval/delivery at page level or minipage level. If it is realized at the page level,
the playout point where it is paused. In other words, the proxy it plays a page and then skips some previous pages [e.g., see
resumes retrieval/delivery at time where is an integer Fig. 11(a)]; if it is realized at the minipage level, it plays a
such that . minipage and then skips some previous minipages [e.g., see
Fig. 11(b)]. In general, if the cycle duration is shorter, the ap-
B. Fast Forward proximation is closer to the ideal one. In particular, the realiza-
Fig. 8(a) shows the ideal fast forward operation. When the tion at the minipage level is better than that at the page level
customer issues a fast forward command, the video program is because a minipage lasts for a shorter duration.
138 IEEE TRANSACTIONS ON MULTIMEDIA, VOL. 5, NO. 1, MARCH 2003
current price is about U.S. $0.2 per yard [17]. Nevertheless, its
bandwidth is about 25 000 GHz [17], and it can currently pro-
vide about one hundred OC-48 channels at 2.5 Gbps/channel
[18], [19]. Even if each channel is operated at the OC-1 rate
50 Mbps/channel (OC-1 is mature and economical), an optical
fiber can still provide a capacity of 100 50 Mbps 5 Gbps.
This is sufficient for the current application (e.g., see the design
(a) (b)
examples in Section V-C).
I/O bandwidth: Given a broadcast delivery scheme, the pages
Fig. 10. Fast rewind operation. (a) Ideal fast rewind operation.
(b) Approximate fast rewind operation. or minipages can be stored in such a way that they are read se-
quentially from the storage for broadcasting. In this manner, the
I/O bottleneck can be avoided. The I/O bandwidth required de-
pends on the number of video programs, but it does not depend
on the number of concurrent customers. Therefore, the proposed
system is particularly suitable for large scale applications with
many customers.
I/O speed and channel bit rate: We can match the I/O speed
of a disk with the bit rate of an optical channel, so that the
system can simply read from a disk for broadcasting over an
optical channel. For example, if we use a disk with I/O speed
50 Mbps [20], we can operate each optical channel at the OC-1
(a)
rate 50 Mbps. In this manner, the system requires an array of
small capacity disks.
MPEG Video: MPEG is a common video compression stan-
dard [21]. It produces three types of frames: I frames, P frames,
and B frames. A group of frames consists of an I frame and a cer-
tain combination of P frames and B frames, and each group is in-
dependent of the other groups. The proposed system can support
MPEG video as follows. Within the VOD warehouse, each mini-
page contains an integral number of groups, and its size is larger
than the average size of these groups by a specified percentage
margin in order to accommodate their variable sizes. Between
the VOD warehouse and the customer, a streaming protocol can
be used to deliver the video to the customer [22].
(b) Video playback rate and duration: In the proposed system,
Fig. 11. Realization of approximate fast rewind operation; the proxy only different video programs can occupy different number of optical
retrieves and delivers the shaded pages. (a) Realization at the page level. channels. Therefore, it can accommodate video programs with
(b) Realization at the minipage level. different playback rate (e.g., 1.5 Mbps for MPEG-1 and 4 Mbps
for MPEG-2) and different duration (e.g., 90 min and 120 min).
V. DESIGN CONSIDERATIONS AND EXAMPLES
B. Selection of Design Parameters
A. Design Issues
Proxy: Each proxy is a simple logical unit for reception and 1) Cycle Duration : The determination of the cycle dura-
transmission. It receives one page per cycle and therefore its tion involves two conflicting factors, as follows.
receiver is utilized in only one of the p slots in each cycle. • If is larger, a channel can broadcast more pages in a
Therefore, multiple proxies can share a receiver. They can also cycle and hence an optical fiber can broadcast more video
share a transmitter for outgoing delivery. For example, if ATM programs. Therefore, a smaller number of optical fibers is
links at 155 Mbps/link are used for outgoing delivery and is needed.
1.5 Mbps, 103 proxies can share a transmitter for one ATM link. • If is larger, the mean access delay is longer. It is because
To simplify implementation, multiple proxies can reside on the when a customer initiates a new VOD session, the associ-
same physical unit for sharing. ated proxy has to wait for an average of s before it
Cost-effectiveness: To serve one additional concurrent cus- can retrieve the first page of the requested video program.
tomer, it is only necessary to add one proxy and it is not nec- The service provider can specify an acceptable mean access
essary to add the other resources. The cost required is small. delay . Then the cycle duration can be chosen to be .
Therefore, the proposed system is more cost-effective for larger 2) Number of Optical Fibers : The number of optical
scale applications with many customers. fibers required depends on the number of video programs,
Optical bandwidth: The proposed system uses an optical fiber the durations of these programs, and the cycle duration. Each
within the VOD warehouse. An optical fiber is cheap, and its optical channel can broadcast bits in a cycle. If the th
LEUNG AND CHAN: DESIGN OF AN INTERACTIVE VOD SYSTEM 139
(1)
than the previous design in Example 1. Consequently, it pro- [14] C. A. Bracket, “Dense wavelength division multiplexing networks: Prin-
vides a better quality (i.e., shorter access delay and better inter- ciples and applications,” IEEE J. Select. Areas Commun., vol. 8, no. 6,
pp. 948–964, Aug. 1990.
active operations) at the expenses of using more optical fibers. [15] M. S. Borella, J. P. Jue, D. Banerjee, B. Ramamurthy, and B. Mukherjee,
“Optical components for WDM lightwave networks,” Proc. IEEE, vol.
85, no. 8, pp. 1274–1307, August 1997.
VI. CONCLUSIONS [16] B. Ozden, R. Rastogi, and A. Silberschatz, “On the design of a low-cost
In this paper, we adopted both the client–server paradigm and video-on-demand storage system,” Multimedia Syst., vol. 4, pp. 40–54,
1996.
the broadcast delivery paradigm to design a VOD system for [17] S. Chatterjee and S. Pawlowski, “All-optical networks,” Commun. ACM,
large scale applications with many customers. This system has vol. 42, no. 6, pp. 74–83, June 1999.
the following desirable features. [18] P. E. Green, “Optical networking update,” IEEE J. Select. Areas
Commun., vol. 14, no. 5, pp. 764–779, June 1996.
• The system can easily be scaled up to serve more concur- [19] G. Wilfong, B. Mikkelsen, C. Doerr, and M. Zirngibl, “WDM
rent customers and provide more video programs. cross-connect architectures with reduced complexity,” IEEE J. Light-
wave Technol., vol. 17, no. 10, pp. 1732–1741, Oct. 1999.
• The system can provide interactive operations which are [20] Seagate Barracuda Disk Drive Product Manual, 1993.
approximations of the ideal ones. The customer can con- [21] B. Furht, “A survey of multimedia compression techniques and
trol the pause duration, the fast forward rate and the fast standards. Part II: Video compression,” Real-Time Imag., vol. 1, pp.
319–337, 1995.
rewind rate. [22] M. Reisslein and K. W. Ross, “High performance prefetching protocols
• The system only involves point-to-point communication for VBR prerecorded video,” IEEE Network, vol. 12, no. 6, pp. 46–55,
between the VOD warehouse and each customer. This type Dec. 1998.
[23] R. K. Ahuja, T. L. Magnanti, and J. B. Orlin, Network Flows: Theory,
of communication can be supported by many existing net- Algorithms and Applications. Englewood Cliffs, NJ: Prentice-Hall,
work infrastructures. 1993, p. 87.
• The system does not involve any network control. This is [24] T. C. Hu, Combinatorial Algorithms. Boston, MA: Addison-Wesley,
1982, pp. 209–222.
important when the network is not owned and managed by [25] A. Scholl, R. Klein, and C. Jurgens, “BISON: A fast hybrid procedure
the VOD service provider. for exactly solving the one-dimensional bin packing problem,” Comput.
• The access delay is small. Oper. Res., vol. 24, no. 7, pp. 627–645, 1997.
[26] G. L. Nemhauser and A. W. Lawrence, Integer and Combinatorial Op-
• Each video stream only requires a small buffer size for timization: Wiley, 1988.
temporary storage. [27] D. Gross and C. M. Harris, Fundamentals of Queueing Theory, 2nd
ed. New York: Wiley, 1985, pp. 294–304.
[28] S. Yakowitz and F. Szidarovszky, An Introduction to Numerical Compu-
REFERENCES tations, 2nd ed. New York: MacMillan, 1989.
[1] D. Deloddere, W. Verbiest, and H. Verhille, “Interactive video on de-
mand,” IEEE Commun. Mag, pp. 82–88, May 1994.
[2] R. Tewari, R. Mukherjee, and D. M. Dias, “Real-time issue for clustered
multimedia servers,”, IBM Res. Rep. RC 20020, 1995.
[3] D. N. Serpanos, L. Georgiadis, and T. Bouloutas, “MMPacking: A load Yiu-Wing Leung (M’92–SM’96) received the B.Sc.
and storage balancing algorithm for distributed multimedia servers,” and Ph.D. degrees from the Chinese University of
IEEE Trans. Circuits Syst. Video Technol, vol. 8, pp. 13–17, Feb. 1998. Hong Kong in 1989 and 1992, respectively. His Ph.D.
[4] D. P. Anderson, “Metascheduling for continuous media,” ACM Trans. advisor was Prof. T. S. Yum.
Comput. Syst., vol. 11, no. 3, pp. 226–252, August 1993. After graduation, he was with the Hong Kong
[5] A. Dan, D. Sitaram, and P. Shahabuddin, “Scheduling policies for an Polytechnic University until 1997. He then joined
on-demand video server with batching,” in Proc. ACM Multimedia Conf. the Hong Kong Baptist University and now he is an
Expo., Oct 1994, pp. 15–23. Associate Professor in the Department of Computer
[6] A. Dan, P. Shahabuddin, D. Sitaram, and D. Towsley, “Channel alloca- Science. He has a wide range of research interests
tion under batching and CVR control in video-on-demand systems,” J. and he has been working on three main areas:
Parallel Distrib. Comput., vol. 30, pp. 168–179, 1995. information networks, multimedia communications
[7] K. C. Almeroth and M. H. Ammar, “The use of multicast delivery to and systems, and cybernetics and systems engineering. He has published
provide a scalable and interactive video-on-demand service,” IEEE J. more than 50 journal papers in these areas, most of which appear in various
Select Areas Commun., vol. 14, no. 6, pp. 1110–1122, Aug. 1996. IEEE publications. He has supervised several Ph.D. students and postdoctoral
[8] W. J. Liao and V. O. K. Li, “The split-and-merge (SAM) protocol for in- fellows, and some of them now teach at various universities in Singapore, the
teractive video-on-demand systems,” in Proc. IEEE INFOCOM, Japan, United Kingdom, Canada, and Hong Kong.
Apr. 1997.
[9] J. W. Wong, “Broadcast delivery,” Proc. IEEE, vol. 76, no. 12, pp.
1566–1577, Dec. 1988.
[10] T. F. Bowen, G. Gopal, G. Herman, T. Hickey, K. C. Lee, W. H. Mans-
field, J. Raitz, and A. Weinrib, “The datacycle architecture,” Commun. Tony K. C. Chan received the M.Sc. degree in
ACM, vol. 35, no. 12, pp. 71–81, 1992. information technology from the Department of
[11] T. Chiueh, “A periodic broadcasting architecture for large-scale residen- Computing, Hong Kong Polytechnic University.
tial video-on-demand service,” Proc. SPIE, vol. 2615, pp. 162–169, Oct. He is currently pursuing the Ph.D. degree in the
1995. Department of Computer Science at Hong Kong
[12] R. O. Banker et al., “Method of providing video on demand with VCR Baptist University. His professional interests include
like functions,” U.S. Patent 5 357 276, 1994. information networks and multimedia systems.
[13] S. Viswanathan and T. Imielinski, “Metropolitan area video-on-demand
service using pyramid broadcasting,” Multimedia Syst., vol. 4, pp.
197–208, 1996.