Traffic throughput between Primary Site and Secondary Sites

[Dario Pattarini]

Dear Experts,
I have a question from a prospect about the throughput generated on the Network between Primary and Secondary.
Do we have any recent data? I know that could be relative, depending on the DB size, but as an average, in order to design properly the Capacity should be fine to have some figures to provide.

The concern is not only about a possible DR design, it is also about how the traffic could impact the Network during their peek times.

I would really appreciate any help or contribution.

Regards,
Dario

[Ralf Prigl]

Hello Dario, There are two scenarios of interest.
(1) Database size
On cluster startup, the full database will have to be retrieved at least by the first node in a secondary site from the primary site. A large PAM database can be several GB in size. If you have a 5GB database and a 100kB transfer speed, it will take 50000 seconds to transfer the file. Or going the other way round, if you wanted this transfer to complete within 500 seconds, you would need a transfer speed of 10MB per second.

(2) User activity
This of course is completely implementation specific, depending on the number of active PAM users, how many sessions they typically have open, and how session recording is configured. If you have local session recording shares in each datacenter, the latter may not come into play. In that case, the bandwidth required for a timely cluster startup, point (1), should be quite adequate for handling ongoing replication activity in the active cluster. This assume that if you have a small database, you most likely don't have a ton of user traffic either.

[Dario Pattarini]

Thanks Ralf. Very much appreciated