On 9/23/2010 9:14 AM, Douglas Hubler wrote: > On Wed, Sep 22, 2010 at 4:43 PM, Talbot, Peter > <[email protected]> wrote: >> In further pursuing a ‘Production Rollout’ ready implementation of sipXecs, >> I had some final questions, mostly in regards to what can and cannot be >> Virtualized. Currently it is looking like anything that actually handles >> voice traffic (whether recorded messages or ongoing conversations) should >> remain physical due to timing constraints, but other Roles can be >> Virtualized. > At cluecon, Anthony M. showed a way to start freeswitch (in sipxecs > this menas AA, and VM) in a mode that attempts to compensate for the > inaccuracy of the timers typical on vms. This showed calls that > slowly degraded in performance as the load increased rather than > choppy after first few calls. With out without this mode, a single > call on a vm always sounded fine and often tricks people into thinking > that there is no difference between vm and physical. > > As a point of reference, on physical hardware, without this timer > compensation, you get more of a hockey-stick curve in degradation > which is obviously the preferred response. > > I cannot seem to find any information on the wiki, but maybe someone > can ask FS mailing list or dig around their archives. I have seen that mentioned here: http://lists.freeswitch.org/pipermail/freeswitch-users/2010-March/054412.html
> On that note, if anyone has sipp load scripts or sipp experience they > can share to build a good platform to analyse this better, please let > me know. > _______________________________________________ > sipx-users mailing list > [email protected] > List Archive: http://list.sipfoundry.org/archive/sipx-users/ _______________________________________________ sipx-users mailing list [email protected] List Archive: http://list.sipfoundry.org/archive/sipx-users/
