Future multimedia-on-demand systems will bring interactive applications in
education, entertainment, advertisement, and visualisation to a large numbe
r of users at homes, offices, classrooms and laboratories. A class of multi
media data, called hypermedia, is represented by graph structures whose bra
nches represent different video and audio segments and whose nodes represen
t user interactions. To provide satisfactory service in an interactive hype
rmedia system, disruptions and delays in the concurrent delivery of continu
ous media data streams from storage devices cannot be tolerated. Slow respo
nse to new requests generated by on-line users is also highly undesirable.
The frequent arrivals and completions of user requests can seriously degrad
e the throughput of typical servers that are designed to maintain performan
ce at the steady state. To address these issues, we propose a dynamic appro
ach to the retrieval of continuous media and sporadic data in an interactiv
e hypermedia server. The throughput and efficiency of our server are improv
ed by observing runtime information to skip or reorder service sequences, a
s well as maximising disk reads. Through pipelining the admission of new st
reams with the departure of ongoing streams, the performance is also immune
to the presence of transients. User response times under identical loads a
re reduced significantly with respect to other existing approaches. Further
more, our scheduler can efficiently redirect disk bandwidth to service spor
adic requests when the demand on real-time services decreases. We evaluate
our technique by comparing its performance with static techniques via thoro
ugh experiments under different buffer capacities, stream lengths and reque
st arrival rates. Results of experiments show that our approach can signifi
cantly improve continuous media data throughput, user response time, as wel
l as sporadic data throughput, and is thus a plausible approach for more in
teractive multimedia applications such as hypermedia.