* Posts by andysparkes

11 posts • joined 28 Sep 2011

There's a lot more to backup than you thought

andysparkes

Its good to see that restore is seen as important as backup and I really liked the comment around the restore times around incremental forever. There is only one backup appliance that I know that actually publishes predicted Restore rates as well as backup ingest rates and that is HP Storeonce. The restore capability was built into the deduping matching and store algorithms to enable that performance. As some of the previous comments suggest, its only when you go to make a claim on your insurance policy that you really know the quality of the policy you bought.

1
0

HP unveils federated backup that won't get your back up ... it hopes

andysparkes

Re: google: Your search - "multilayed hashing schema" - did not match any documents.

I can assure you that I'm not a marketing guy.... I guess what I was trying to say was that you can hash at one level of chunks and then a combination of chunks or their associated metadata can also be hashed. Techniques like this can be used to provide another layer of protection against any theoretical hash collision.

1
0
andysparkes

Re: global dedupe

Storeonce uses a multi-layed hashing schema that means the chance of hash collision is practically zero. As well as a SHA-1 hash other mechanisms are used to perform a series of checks so that what is hashed and written can be recovered.

Theoretically a Storeonce Dedupe store could span across 1.7PB and even with those numbers of hashes the maths and probability is on our side by a substantial margin.

2
1
andysparkes

Federated catalyst does allow dedupe across a number of stores but its power is in the virtualization of stores, a single store can be created that can be subsequently extended vastly increasing the flexibility and simplicity of management, The virtualization also means that ultimately the stores can exist anywhere. Together with the HA capabilities of the 6500, its certainly true that HP is taking a different tack than EMC, its solving the problem of a series of point products with a seamless, scalable and virtualized software architecture.

1
0

How to grill an EMC division chief

andysparkes

Re: pancakes

Well, storeonce does dedupe across nodes. Catalyst stores can be created across nodes and these can scale across 4 nodes and across all the capacity in the node if you so desired. Instead of having to wheel in yet another data domain box, go through the grief of setting up more stores, HP federated catalyst allows you to extend your stores as you grow. HP Discover 2014.

2
0
andysparkes

Well, HP storeonce is a scale out dedupe appliane. What dedupe across stores gives you is the ability to add and extend your existing stores without having to create new stores as you grow... you also get the supplementary benefit of increasing data set you can dedupe against. HP storeonce is already available as both a physical and virtual appliance.

2
0

We trust computers to fly jets... why not trust them with our petabytes?

andysparkes

Actually there are some orchestration engines that do allow storage to be added as a set of parameters and then deployed against application needs.... HP Orchestration Operations is one of them. It enables the automatic deployment of servers/storage for applications and VMs

0
0

HP boffin: Honey! I shrank the PC. To nanometre size, dammit

andysparkes

the boffins will win, or more specifically HP will win.... HP IS returning to its engineering roots and still has a very deep bench of talent. The enterprise group is run by an engineer and HP Labs is also run by engineers.

3
0
andysparkes

There are parts of HP that are extremely innovative, next weeks HP Discover will also highlight some very smart integration of HP Labs IP into HP Storage products. There will also be demonstrations of how Autonomy can work inconjunction with HP's storage products

2
0

EMC Isilon stiffens penetration into NAS arena

andysparkes

X9000

HP X9000 products also a real contender in the scale out NAS space.

0
0

HPC 2.0: The Monster Mash-up

andysparkes

it certainly is a storage challenge....

Just storing the stuff on disks seems easy doesn't it... But look at what's happening to bandwidth verse capacity growth - it cannot keep up. It means that you really will see the compute move to where the data is being generated, Perhaps filesystems and storage arrays need to get a lot smarter than they are today.

0
0

Forums