The first episode of Raising the Floor, the enterprise datacenter podcast from Foskett Services, features thought leaders from two key cloud storage-enabling companies:
We conclude our conversation by confronting some of the objections people have to cloud storage: Security, data protection, and performance.
Stephen: Now, not everybody’s onboard though obviously. There are easy objections that people have beyond cost. Cost is probably the first objection that people have but the second is security. There, I think that people are maybe looking at the wrong thing. I hear a lot that they’re worried about providers going out of business or having an outage or losing their data.
To me, I think a greater concern is what might happen to that data once it’s outside the walls of the datacenter. How do you answer that? Let’s start with Nasuni, how do you answer the security concern?
Andres: I think this fear goes back to the beginning of our conversation. Which is, one of the great benefits of the cloud is that it can be a multi-tenet environment. There are many people sharing the same infrastructure. This fear of data leakage or essentially what’s going to happen because I’m sharing infrastructure with other people. If I put my Social Security numbers of my employees up there in the cloud, can someone else hack into that and look at those Social Security numbers? That is the fear.
The fear is that private business information will somehow leak into the Internet because that’s where it’s being store. The way we like to address this is the cloud is for storage. A heck of a lot more secure than it’s ever been for either software as a service. Things like SalesForce.com or Cloud Compute.
That’s simply because, when your data is in your datacenter and when you have products like the Cirtas product or the Nasuni products, everything gets encrypted on your floor with your own encryption keys. Then, it leaves the datacenter encrypted already.
Those encryption keys are never seen by Nasuni. They do not have to be seen by your cloud provider. Therefore, your data in the cloud will be encrypted to this kind of top secret — it takes a million computers a million years to hack into AES 256 which is what we all like to use for this crypto, out there in the cloud.
Really, the fact that neither the provider nor the vendor, like we are, has to have your encryption keys give you an enormous edge. When you have all our customer information out there on SalesForce.com, their servers need to be able to see your data.
SalesForce has put in place very many good procedures to protect your data, but it’s procedure-based. When we do it, it’s really crypto-based. It’s the encryption itself is protecting your data. Because, no one else needs to see your data out there, to be able to make use of it.
So, cloud storage is very, very secure. Much more secure than things we’ve already been using for 10 years.
Stephen: I think that the other thing people are concerned about is performance. I think that this affects block storage especially, Josh. File applications are a little bit more tolerant of latency and the time it requires to access data. Block storage applications are much less tolerant of latency. Latency, of course, is the plague of the Internet. How do you address the latency question?
Josh: There’s several techniques that we use in our products to do this. Just by way of background, the technical heritage of a lot of the engineers who started our company comes from the WAN acceleration space. These are guys who have tackled this problem many times before and made lots of things work well over the Internet that were never designed to work over networks. It’s a core competency for our organization.
Specifically, things that we’re doing are protocol optimization out to the cloud. The de-duplication and compression algorithms that we use are part of this as well. By taking data and getting it down into as small a format as possible, you’re moving less information around so you can do it faster.
Finally, our system is delivered as a hardware based appliance. There’s a sizeable caching system on that that spans memory and SSD as well as spinning disks. What the system is doing is always looking at the data that you’re accessing and keeping those working sets of data local to the appliance where we can serve them with SAN levels of latency.
There’s only the occasional miss, a piece of information we do not have local on the appliance that we have to go and fetch from the cloud. That’s minimized quite a bit and we do this in our sizing of the caching system on the appliance to the workloads that we’re targeting.
Stephen: That’s great. I think that clearly the WAN acceleration market has demonstrated that it is possible to do things over the Internet that were seemingly impossible. The history of the last 10 years has demonstrated that it’s possible to do things over the Internet that nobody else thought possible. Here we are recording a call like this. I guess one final thing, one final thought. Is cloud storage a viable primary storage media or is it really, as people see it, as an alternative or secondary storage media. Is this a primary storage media?
You want to jump in there Andres?
Andres: I think this is a question of capability versus what people are comfortable with today. What Josh just talked about, is in my opinion, what defines this new generation of gateway or Ray products that are connecting the cloud to the datacenter. Here in Nasuni, the team here comes from the EMC symmetric’s caching side of the business. It’s all the same kind of thing where you have a lot of high-performance caching. Understanding how to cache a file system, such that, you can achieve the same thing that Josh was just talking about which is give me primarily performance in a way that I can make it indistinguishable from what I’m using today.
If you don’t do that then, it gets complicated. I now have to think of what is appropriate to tier to the cloud and what I can keep in my data center. I’m not claiming that businesses in general are ready to put everything out there on the cloud, yet. The capability’s there.
I think what we lack a lot of is not just comfort but, awareness. People have been using the cloud for online backup for many years now, or cloud-like systems, like Mozy and Carbonite and things like that.
Now, you can go one step closer to something that’s really, really useful. You can put a primary file server and you can connect your users on it. Your users will not know that they’re taking it from the cloud.
You can bring in something like Exchange or something like SharePoint and plug it in front of a Cirtas box and the users won’t know the difference. That is the goal of this new generation of connector products, un-ramping products to the cloud.
It’s really to blur the distinction between cloud users and local users for storage.
Stephen: I think that’s really exciting. I want to thank you guys for joining me today for this “Raising the Floor” podcast. Josh Goldstein from Cirtas, thank you very much for joining us.
Josh: Thank you Stephen, and thank you Andres, it was a pleasure.
Stephen: Yes, and thank you Andres, Andres Rodriguez from Nasuni.