Veeam and StoreOnce capacity calculator?

[b.vanhaastrecht]

Hello,

All to us known HP calculators work with traditional GFS scheme's, and does not handle Veeam's way of creating backup chains.

Since there is an approved by Veeam stamp on the StoreOnce boxes, is there an Veeam & StoreOnce capacity calculator somewere?

Kind Regards,
Bastiaan

[Dima P.]

Hello Bastiaan,

Have not seen any. This might be a silly question but can you share the link to the existing HP calculators? Thank you in advance.

[b.vanhaastrecht]

Thoose tools are from the HP partner portal and are NDA. So if you are an HP partner (which I presume Veeam is) you can get them there. Its called storage sizer tools.

We got to make an sizing for a customer with a huge dataset. We need to estimate which HP StoreOnce base device we need to offer. Can't imagine Veeam hasnt got any experiance since you've got an alliance with HP and the StoreOnce boxes. Where to go from gere?

[dellock6]

Bastiaan,
while I can't reply about the HP partnership and their tools, if you are going to store regular chains in the storeonce, the dedupe benefits (which are the only difference compared to a non-deduped storage in terms of space usage) will heavily depend on what kind of configuration you will have in the backup jobs.
Assuming the daily incrementals are hardly deduped as they contain unique data, the dedupe results will come from the fulls you create, and for those I'd suppose some 80-90% of reduction could be possible as they contain the same VMs (but at different point in time).

That said, I'd start by using the size calculator from Timo: http://rps.dewin.me/ and then remove the said 80-90% of the space consumed by the additional fulls. And probably I'd stay on the safe side and use a lower numbeer than 80%. Deduplication cannot be easily predicted...

[Gostev]

If deduplication was predictable, sizing tools from dedupe vendors would be available publicly

[dellock6]

Hardware deduplication is usually "up to 50x*" with the * pointing to a font4 footer note that states, among other things, that you can reach that result if you run dedupe in an odd day, during a full moon, chanting pagan songs, and sacrificing a goat on the altar of marketing numbers

[b.vanhaastrecht]

It's actually a great tool, and the download is publicly, the license to run it however is only provided to HPE Partners (because they get trained to operate it).

Software: http://h30144.www3.hp.com/SWDSizerWeb/
License: https://partner.hpe.com/group/upp-emea/partner-tools

We and our fellow HPE partners are using it very frequent to size HPE storage solutions. And it has proved to be very accurate, HPE has loads of reference data to make accurate calculations. You say dedupe is unpredictable, well the tool has multiple data types to select from (SQL/Exchange/File/Image/Oracle…) and a mass of other options to match it to the data type, retention, backup scheme and so on. And you can select which backup software is being used. And I was surprised, dispite on Veeam's Alliance with HPE, it isn't there to select. (Yes there are other 3'rd party backup software’s in there).

So instead of hiding after ‘it’s unpredictable’ , get on with HPE and add Veeam to the Storage Sizer Tool. At least we have some reference instead of (educated) guesses.

[Gostev]

Got it. In that case, we will certainly reach out to our HP contacts to see how we can go about that. Thanks for the info!

[daddolanglois]

Depends if the solution is using StoreOnce Catalyst (i.e. Veeam v9). Having sold HP StoreOnce & Catalyst solutions in a previous role, its fair to say that I've seen dedupe ratios with StoreOnce for VM's of between 24:1 through to in excess of 50:1. Check out the video on youtube.com from Federico Venier from VeeamOn 2015 - his lab tests show a 15:1 dedupe ratio, which is probably a very safe bet.
https://www.youtube.com/watch?v=fKRzv0lcKpg

[mkaec]

Hardware deduplication is usually "up to 50x*" with the * pointing to a font4 footer note that states, among other things, that you can reach that result if you run dedupe in an odd day, during a full moon, chanting pagan songs, and sacrificing a goat on the altar of marketing numbers

And also run full backups every day. (Or more for an even greater ratio!)

To go slightly off-topic, the sales reps for Nimble gave us compression stats from actual usage across all of their clients. That felt more reliable than the typical "we achieved this in a lab" stats. If you've got several thousand users all achieving a certain compression ration on SQL data, then I'd say that's a pretty reliable indicator for the next customer.