Human-Computer Interaction 3e Dix, Finlay, Abowd, Beale
EXERCISE 19.19 [extra - not in book]
Site X on a network has connections to 4 other sites A, B, C and D
The connections have the following characteristics:
site X â€“ site A high bandwidth low latency low jitter
site X â€“ site B high bandwidth high latency low jitter
site X â€“ site C low bandwidth low latency low jitter
site X â€“ site D high bandwidth low latency high jitter
(a) audio-conference (e.g. telephone style chatting)
(b) live interactive video-only connection (e.g. web cam)
(c) streaming HI-FI audio (e.g. Radio 3)
(d) video-on-demand (e.g. hotel pay-per-view video service)
For this part assume that there is not enough memory available to provide any buffering. Make sure your answer clearly states the effects on the users, not just technical effects.
Two potential architectural solutions have been proposed
Describe advantages and disadvantages of these solutions with regard to potential deployment on:
answer available for tutors only
This answer may be organised as media within sites (as here) the other way round, or a mix.
Xâ€“A â€“ high bandwidth, low delays and low jitter â€“ good for everything!
Xâ€“B â€“ high latency will behave badly with both audio and video for interactive conversations (a) & (b). In particular, normal conversation has very small gaps (fraction of a second) which are used by participants to hand-over to one another. The participants will have problems similar to trans-continental satellite links. The high latency will not hinder streaming video/audio which can simply be slightly delayed.
Xâ€“C â€“ low bandwidth â€“ will mean high compression and/or loss of quality. This will rule it out for Hi-Fi audio and high quality video. Interactive audio may cope (depending on how low the bandwidth) and web cam can simply have reduced frame rate and resuloution.
Xâ€“D â€“ high jitter â€“ (without buffering) will mean that both interactive and streaming audio will be badly affected (we can not stand delays/missing parts of speech). The interactive video will simply occasionally have broken frames which is acceptable. Depending on the level of jitter this will probably mean the quality is too bad for hotel video application.
In case Xâ€“B, the applet based solution would be necessary to achieve good interactive feedback for free-hand drawing. However, applet security would mean that the server would need to sit on the same machine as the web server (or need signed applet). Also the software will need to be more complicated as there will be the possibilities of race conditions when updates are done simultaneously by different participants. Feedthrough will also be poor.
In case Xâ€“C, the central server solution may be chosen as the low latency and jitter mean that it can give fully interactive feedback even for mouse movements (ideally students may quote figures of around 200 ms maximum round trip feedback time for this). This will have an easier implementation. There may still be problems however in updating photo images on the individual displays, perhaps a delay when the application is first launched as the photos are downloaded over the low bandwidth connection. Feedthrough better.
Other exercises in this chapter
ex.19.1 (ans), ex.19.2 (ans), ex.19.3 (tut), ex.19.4 (tut), ex.19.5 (tut), ex.19.6 (tut), ex.19.7 (tut), ex.19.8 (tut), ex.19.9 (tut), ex.19.10 (tut), ex.19.11 (tut), ex.19.12 (tut), ex.19.13 (tut), ex.19.14 (tut), ex.19.15 (tut), ex.19.16 (tut), ex.19.17 (tut), ex.19.18 (tut), ex.19.19 (tut), ex.19.20 (tut)
all exercises for this chapter