Conclusion
We hope that Web caches will replace all archives and mirrors on the
net.
Theoretical work to be done:
- construct mathematical models of caching systems
- how to set expiration times?
- how cheap must net bandwidth be to make pre-fetching/replication worthwhile?
Experimental work to be done:
- construct extremely large cache hierarchies, generate log data about
page accesses
- make log data publicly available for research