[ http://issues.apache.org/jira/browse/NUTCH-180?page=comments#action_12362816 ]
Andrzej Bialecki commented on NUTCH-180: ----------------------------------------- The "big three" use an efficient caching + a dozen other tricks. Please see the discussion on the IndexSorter on nutch-dev mailing list. > Performance problem with widely used keywords > --------------------------------------------- > > Key: NUTCH-180 > URL: http://issues.apache.org/jira/browse/NUTCH-180 > Project: Nutch > Type: Wish > Components: searcher > Reporter: Mike Alulin > > It looks like Nutch is very slow when the search phrase includes a few widely > used keywords. For example "I 1 2 3 4 5 6 7 8 9 0" typed without the quotes > to Yahoo, Google, or MSN is processed in less than a second. Nutch on the > other hand requires much more time for this even on smaller databases. For > example this phrase made objectssearch.com think more than 1 minute although > their DB is much smaller than DBs of the big 3 guys. On my test Nutch DB with > only 3M pages this phrase took a few seconds to process. > Unfortunately I do not know much about search algorithms, but it looks like > Nutch do have some space to improve the search performance. The current > implementation can be easily "killed" by a few search requests like this. > Just a couple of dozen of such requests makes my server with 2 Opterons think > for a minute or two with 100% CPU utilization. -- This message is automatically generated by JIRA. - If you think it was sent incorrectly contact one of the administrators: http://issues.apache.org/jira/secure/Administrators.jspa - For more information on JIRA, see: http://www.atlassian.com/software/jira
