Feature request: Backup storage sizing feature

[axpett]

It seems that Veeam dosen't provide a backup storage sizing tool.

There are of course a few helpful web sites (e.g. http://rps.dewin.me/), but all of them do their calcualtions based on a total # of TB to backup, and estimations regarding dedup & compression only.
What's needed, is an assessment tool that provides reliable results, based on the customer's real data to backup, i.e. data quantity, file types, achievable compression, dedup ratio, ...

I've been thinking of a solution for quite some time, and just recently had an idea.

Wouldn't it be perfect to use Veeam Backup & Recovery itself for this?

What I'm thinking of, is to run B&R in some kind of assessment mode, in which it runs the backup jobs as usual (read source data, compress, dedup), but only stores the required metadata (not the user data) that's needed for subsequent differential, or incremental backups.
With such a feature, one could configure B&R as required, and get the wanted sizing information from e.g. the log files, or reports.
To activate that feature, I'm thinking af a special assessment license file with a sufficient amount of socket, and instance licenses (unlimited would be perfect).
Once the assessment is done, and a properly sized backup repository is in place, all that the customer needs to do is to replace the license file.

Does this sound reasonable? Thoughts?

André

[HannesK]

Hello,
and welcome to the forums.

The idea is good and not new. There was a tool out there, but as far as I know, it doesn't work with the latest versions of VMware. I also know that some guys are working on a new version of that tool, but it is in an early stage.

Yes, VBR would be a good source, but not "perfect". One thing is that midsize+ customers do not allow to take snapshots of all their VMs. And that's the main problem. With snapshots & CBT data, one could calculate. But it requires snapshots. Assuming that a customer allows to take snapshots of all his VMs, your idea with the "only store metadata" is quite creative, yes For me, this is still too complex. I usually had customers with several (sometimes hundreds) branch offices. So all these VMs somehow need to go through VBR which is quite complex to configure "just for data reduction estimation".

The fun fact is: using 50% data reduction and 5% change rate (up to 100TB source) and 3% (above 100TB source) worked really good in my 4 years as Veeam SE. And nobody can predict the future anyway. If the customer decides to deploy a 50TB encrypted VM (encryption inside a VM which results in 0% data reduction) in a 200TB project, then he will be in trouble no matter which sizing tool you used before (I have been in such a project).

Best regards,
Hannes

[axpett]

Hannes,

thanks for the quick feedback, and for sharing your experience.

I do agree that an estimation based on past experiences may work for a lot of environments, and that there are always several varaibles that need to be taken into account, which can only be estimated.
However, in environments where such estimations result in hundreds of TB required backup storage, I really need a closer, more realistic sizing, which will - based on the results - help deciding whether e.g. some large JBOD with Veeam's software based dedup and compression, or a dedup appliance with global dedup and compression (with all the pros and cons) will be a better solution.

If a customer is interested in a new backup solution (convincing someone to replace the backup solution is one of the toughest things), I don't see why he wouldn't agree to run an assessment. This certainly needs to be properly planned.
Regarding your branch offices example, it may certainly not be required to run this on the whole environment, if it can be savely assumed that the branch offices have similar data.

Please don't get me wrong, I don't expect something that will allow me to switch off my brain, and strictly rely on the tool's results, but IMO a proper sizing requires more than just estimations.

André

[HannesK]

Hi,
as said, your request makes sense, a tool would be nice. But even for hundreds of TBs or PBs, no tool can really predict future. Nobody can guarantee that what you test today is still valid in some month when you order hardware. IT changes too fast.

For dedupe appliance vs. normal backup: The deduplication appliance vendors should have their own methods of prediction. I would not use Veeam methods for that.

in general: never use dedupe appliances Well, if you do active full backups very often, there might be use cases, but even then, keep in mind that restore performance is not good compared to normal storage. So even if you buy 2x or 3x the space of a dedupe appliance, it is still cheaper and faster with a normal server (I try to save money, not disk space). That's why I recommend the rule of thumb. I understand your concerns, I worked a lot with partners, but after some time they see that the rule of thumb worked and no scientific sizing is required.

Best regards,
Hannes