Conclusion
In this work, we have explained the details of our client based
prefetching scheme. We have mentioned other research into this area,
but here we uniquely provide working code that may be analyzed so that
it's performance can be evaluated in real world, as opposed to
simulated, conditions. We have also tried to make our implementation
as open as possible, not only so that it's limitations can be removed
in the future, but so that it can be modified and expanded upon as new
situations arise. It is our hope that this initial working backbone
can be expanded upon to produce a fully implemented
high performance Web browser. During the course of this paper, the
limitations of the current code have been pointed out, and this is the
beginning of future work. Also, there are numerous small
modifications that could be implemented to improve performance even
more should the analysis of the data prove that prefetching will make
a significant performance contribution. To take prefetching one step
further, server side prefetching has already been proposed, so a
logical next step is to implement that aspect, and consider
the possibilities in implementing client/server cooperative
prefetching schemes.
Back |
Index |
Forward