[Seagull-users] Memory requirement per diameter data session

[Iqtidar R. <ik...@gm...>]

Hi,

Does anyone have experience trying to simulate millions of concurrent data
sessions using diameter with Seagull?

I am interested in finding out the memory per session required by Seagull
to maintain the sessions; as typically data sessions can be fairly long.

e.g. 300 session starts / second with hold time of 3600 seconds will result
in peak of approx. million concurrent sessions.

Need this info to select the right hardware for performance testing.

Please also share if any limitations are known for max. concurrent data
sessions support.

Thanks,

Iktidar