Next: About this document
Up: Simulation Result
Previous: Half Data Sharing
- For each of the algorithms, the throughput increases as the degree
of data sharing increases.
- For query mixes M1 and M2, the efficiencies of the different
algorithms were close. Because every query accessed the same copy of the
database, it was easy for any algorithm to keep the important portion
of any database in memory. But RAND and FIFO performed slightly
worse than other algorithms, due to their inherent deficiency in
capturing locality of reference.
- For query mix M3 the performance of different algorithms diverged.
M3 contains more of small queries. The "working" portion of the database
becomes less distinct as many small queries enter and leave the system.
Therefore, algorithms that made an effort to identify the localities
performed better than those that did not.
Deepak Kumar Tawri
Wed Apr 21 21:25:41 IST 1999