Highlighted

Disaster recovery hardware

Hello, my question is, if I have a master server with 32TB in disk deduplication and 48 GB of ram connected to an MSL2024 library, for the server that will do the disaster recovery do I need the same amount of disk and ram?

 

Thanks for the answers

1 Solution

Accepted Solutions
Accepted Solution!

Re: Disaster recovery hardware

From an initial simplistic (my own) viewpoint, then my own personal immediate answer is yes, assuming that you need a quick answer to a quick question, because in a DR situation that same "infrastructure" will most probably be expected to also then perform the same range of backups and retention and restores - assuming, of course, that the original was sized correctly for backups and retention and restores at the primary production site.

One could actually argue that the DR "backup" infrastructure might actually need to be more powerful, more capable, because... it needs to not be a "DR backup" platform it actually needs to be a "DR restore" platform because... modern dedupe backup software is more weighted towards ingest, and less weighted towatrds egress - e.g. one might deploy a platform (masters and/or medias) that can peform an effective (after accelerator and client-side dedupe) ingress of for example 1TB/hr in a 14hr backup window... but to achieve a restore capability of 1TB/hr in a 14hr restore window is most likely a whole different proposition.

So, at the end of day... go back to your RPO and RTO as these are your service commitments... then once those are established, look at your data volumes, platform complexities, and business application service levels (which first, which next, which last)... which determine your required data rehydration and data transfer rates... which determines your hardware and software requirements... and it is only this full "awareness" of actual DR "restore demand" that will inform and determine your actual DR platform requirements.

View solution in original post

1 Reply
Accepted Solution!

Re: Disaster recovery hardware

From an initial simplistic (my own) viewpoint, then my own personal immediate answer is yes, assuming that you need a quick answer to a quick question, because in a DR situation that same "infrastructure" will most probably be expected to also then perform the same range of backups and retention and restores - assuming, of course, that the original was sized correctly for backups and retention and restores at the primary production site.

One could actually argue that the DR "backup" infrastructure might actually need to be more powerful, more capable, because... it needs to not be a "DR backup" platform it actually needs to be a "DR restore" platform because... modern dedupe backup software is more weighted towards ingest, and less weighted towatrds egress - e.g. one might deploy a platform (masters and/or medias) that can peform an effective (after accelerator and client-side dedupe) ingress of for example 1TB/hr in a 14hr backup window... but to achieve a restore capability of 1TB/hr in a 14hr restore window is most likely a whole different proposition.

So, at the end of day... go back to your RPO and RTO as these are your service commitments... then once those are established, look at your data volumes, platform complexities, and business application service levels (which first, which next, which last)... which determine your required data rehydration and data transfer rates... which determines your hardware and software requirements... and it is only this full "awareness" of actual DR "restore demand" that will inform and determine your actual DR platform requirements.

View solution in original post