tag:blogger.com,1999:blog-2954359812249072053.post1271986464146309612..comments2023-12-15T23:33:59.034-06:00Comments on Cary Millsap: RobB's Question about M/M/mCary Millsaphttp://www.blogger.com/profile/16697498718050285274noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-2954359812249072053.post-6077648576718080742010-08-13T15:12:58.413-05:002010-08-13T15:12:58.413-05:00Cary,
Practical Queueing Analysis (Mike Tanner) -...Cary,<br /><br />Practical Queueing Analysis (Mike Tanner) -- has some good discussion of service time variance (0 = constant service time) for single-server queues (M/G/1) (page 95) and for multi-server queues (G/G/1) (page 217). Generally, variability in service time increases average waiting time...Dominic Delmolinohttps://www.blogger.com/profile/04315832976299102261noreply@blogger.comtag:blogger.com,1999:blog-2954359812249072053.post-29817426481569366642010-03-06T02:36:17.187-06:002010-03-06T02:36:17.187-06:00Metasoft,
Indeed; thank you for pointing that out...Metasoft,<br /><br />Indeed; thank you for pointing that out. I agree with you there. That worst-case observation is consistent with the manner in which I recommend applying the model. <br /><br />--CaryCary Millsaphttps://www.blogger.com/profile/16697498718050285274noreply@blogger.comtag:blogger.com,1999:blog-2954359812249072053.post-38394427144860215492010-03-05T22:33:18.655-06:002010-03-05T22:33:18.655-06:00hi cary,
just a thought.
hope this doesn't m...hi cary,<br /><br />just a thought.<br /><br />hope this doesn't mislead you down the wrong path.<br /><br />M/M/m is open queue. Typical DB workload is closed queue. So M/M/m is lower bound on performance, or worst case performance.metasofthttps://www.blogger.com/profile/17149213781391733478noreply@blogger.com