Nasuni and cloud storage

I had a chance to meet with the CEO from Nasuni yesterday, Andres Rodriguez (@
nasunista), and get a look at what they are doing. Wow Cool stuff. We will work up a cloudscore for them soon but they said all the right things. 

But first let me back up. A few months ago at VTUG's Fall Forward event I got to hear Steve Duplessie (@stevedupe) from ESG talk about the future of storage. One of the things that stuck with me was his idea that virtual storage controllers were the next big storage thing. I didn't quite get it then, some times it takes a little time to roll around my brain before it settles in, but after looking at Nasuni I get it.

Normally storage controllers are tied to the physical storage, but imagine if they were not. Nasuni essentially builds a controller (and NFS,CIFS, iSCSI front end) that is independent to the storage. In fact the back end storage they use now is Amazon S3.

They have a local cache to get rid of the performance concerns but all the storage, replication, backups etc are done with S3. In fact someone used the analogy (and I love analogies) that it is like EMC + Akamai.

They encrypt the data and the customer has the keys so compliance and security should not be a concern. S3 is super reliable when configured correctly and if the on site box crashes you can configure a new virtual one in 15 minutes.

The other interesting thing that they are not doing yet, but probably can, is connect to other types of storage. Imagine, if you had box, amazon, dropbox and local storage all managed through this one virtual controller. Also, since we are making stuff up, that you can scale multiple ones of these together to get super high  performance, only limited by the network connection. With 100Gb Ethernet on the radar that's a pretty good size data pipe. String 10 of these in parallel you could get 1Tb access speeds to unlimited data. Talk about big data....

The other thing that they do that is pretty unique is you pay only for usable storage, not usage, not snapshots or number of users. This makes it really easy to understand the cost. It is literally COST X  USABLESTORAGE.

When comparing costs though you do need to factor in some of the other benefits you get. Disaster recovery is built in. No need to replicate data, or buy extra hardware, you can literally configure a VM appliance in minutes, or have it already running in the second (or third) sites. You get centralized storage in the cloud with data backups for free too. Which saves a lot of compliance headaches and missed tape pickups..

Still working on the math to see if this is cheaper or not than local storage, plus backup licenses and media, plus a hot DR site plus the time to manage it plus the WAN link etc... I'm thinking it will be pretty close.

Cloud gives the ultimate in flexibility

In today's world it's important that companies be nimble and flexible. Things change too quick to be stuck with a rigid infrastructure, processes or mindset.

This is where cloud, in particular SAAS, really come into play. In my opinion cloud for cost savings is short sighted, not that there can't be cost savings, but that should not be the reason to go cloud. It's really about being faster and more flexible.

Imagine two companies, one with on premise infrastructure and one cloud based. Both want to merge with another company and add 1000 "new" employees to their email system because of it. The on-premise company needs to upgrade it's email servers, storage, add backup licenses and add the new users. Let's assume they are all outlook users and don't need training since it's the same interface. Still adding servers and storage (and possible network ports) can take a month and cost $50,000. If disaster recovery is important (and it should be) you may need to double that.

Now the cloud based company simply process a change order (or whatever paperwork is required  to add 1000 new users and add's their accounts. This can literally be done in a week. Technically hours, but there is usually some level of financial approval required for an additional monthly cost like that.

The other way cloud makes you more flexible is the fact that upgrades can happen with no real IT work needed. Now IT needs to spend time on testing the new features, figuring out use cases and then training the user community on the best way to leverage them, but they should do that for on-premise features too.