kea-dhcp4 - benchmarks (memfile and mysql)

Chaigneau, Nicolas nicolas.chaigneau at capgemini.com
Wed Sep 17 08:58:15 UTC 2014


> Hello,
> 
> 
> Some update (and very good news).
> 
> I've run again the benchs with MySQL, with only one parameter changed, as you suggested:
> 
> innodb_flush_log_at_trx_commit = 2
> 
> (instead of the default of 1)
> 
> 
> 
> This changes *everything*.
> 
> I'm now able to serve 800 leases/s (instead of 50, with the default value of innodb_flush_log_at_trx_commit).
> As compared to 4900 with a memfile with lease file on RAMDISK.


4100, not 4900 (4900 is without writing the leases to disk).



> (And 70 for dhcpd with lease file on RAMDISK).
> 
> This is a simple preliminary bench scenario with only DORA for new clients, but still valid as a comparison standpoint.
> And it's promising :)
> 
> 
> 
> I understand that changing this parameter entails that we're not "ACID" anymore, and can lose about 1 second of data in case of a crash.
> This means losing at most 800 leases, which is acceptable.
> 
> 
> 
> Thanks again for your help.
> 
> 
> Regards,
> Nicolas.
>
This message contains information that may be privileged or confidential and is the property of the Capgemini Group. It is intended only for the person to whom it is addressed. If you are not the intended recipient, you are not authorized to read, print, retain, copy, disseminate, distribute, or use this message or any part thereof. If you receive this message in error, please notify the sender immediately and delete all copies of this message.



More information about the kea-dev mailing list