ssh does encryption, and encryption is a type of minor compression, so you may be transmitting less data on the line.
On Monday, February 18, 2013 3:15:07 PM UTC-7, Richard wrote: > > Final result : > > no_ssh : 32.810 seconds (worst) > remote with ssh : 32.010 seconds > local : 31.494 seconds > > These are pretty basic tests, with profiler my dev machine may not be in > exact same state (more memory or less for the differents test particularly > the last test with no ssh). > > But it pretty strange that no ssh take even longer then with ssh... > > Richard > > > > Le lundi 18 février 2013 16:49:30 UTC-5, Niphlod a écrit : >> >> well, it depends. >> distinct boxes: more latency for every connection >> one-box: less latency but possibly ram,cpu,disk contention >> If your app uses a lot of CPU (that doesn't mean "your app do expensive >> db queries, that cpu is used by the db process") or a lot of memory or does >> a lot of writing/reading from disks, then having it separate can be better >> then having all in one box. response.toolbar() holds all the timings of the >> last queries, so you can use it to time the difference between a query on >> the local db instance and on the remote one. >> >> PS: ssh encryption on all traffic from/to the database adds another layer. >> >> On Monday, February 18, 2013 10:43:24 PM UTC+1, Richard wrote: >>> >>> Hello, >>> >>> I would like to know if someone has any experience to report here about >>> separating web and database server into distinct server (2 linux box for >>> example). >>> >>> I making some test actually and found no great improvement to do so... I >>> access the remote database server through ssh port fowarding, so it may be >>> the reason why I have no speed improvement. >>> >>> But I read that if web app is making a lot of db request (1000+) it may >>> be a bad idea to split web server and database server ( >>> http://www.mysqlperformanceblog.com/2006/10/16/should-mysql-and-web-server-share-the-same-box/). >>> >>> In web2py "represent=lambda..." trigger a lot of db request, so I wonder if >>> web2py represent could be responsible of the no speed improvement. >>> >>> I use profiler and the difference between local and remote database is >>> pretty small (.5 sec longer with remote db). >>> >>> I will try to profile remote database with no ssh to see if I get speed >>> improvement and get back here. >>> >>> Thanks >>> >>> Richard >>> >>> PS.: I try to use "performance" as a tag, but I can't add new tag, but I >>> think that having a "performance" tag or "speed" tag could be great... >>> >>> >>> >>> >>> -- --- You received this message because you are subscribed to the Google Groups "web2py-users" group. To unsubscribe from this group and stop receiving emails from it, send an email to [email protected]. For more options, visit https://groups.google.com/groups/opt_out.
