This article is part of an Essential Guide, our editor-selected collection of our best articles, videos and other content on this topic. Explore more in this guide:
2. - Making smart cloud storage controller choices: Read more in this section
- Tips for buying a cloud storage appliance
- Ways to evaluate a cloud controller
- Cloud gateway use cases
Explore other sections in this guide:
Can cloud gateways compete on technical merits? How much consolidation will occur in this space? Arun Taneja, founder and consulting analyst at the Taneja Group, answers key questions about cloud gateway appliances and how they factor into disaster recovery and primary storage use cases. Listen to the podcast he recorded with Executive Editor Ellen O'Brien, or read the transcript below to learn more.
A cloud gateway appliance is a separate purchase, but will they evolve into built-in offerings, ones that are actually part of your array?
Arun Taneja: The short answer is absolutely they will -- but I have to say that none of us should hold our breath for that to happen. Let's look at the dynamics in the marketplace. A large majority of the storage is provided to the enterprise and small and medium-sized companies, from the likes of EMC and IBM, HP, Dell, NetApp, and Hitachi. And they're all in the business of selling disk drives. A large amount of storage means good revenue for them.
Gateways of these types essentially make the cloud the storage element; their impetus to make this gateway solution happen quickly, [is] almost zero. But having said that, I know that when it's a good idea, it always comes to pass.
When you look at the newest features available in gateways, which are the most important?
Taneja: I absolutely agree that cloud gateways are definitely starting to mature. I would say that as a market, we are sort of the third incarnation at this stage of the game. Now, which features are the most important? I would say you almost have to ask yourself, 'What functionality is this gateway primarily designed to deliver?' because there's a variety of them in the marketplace, some that are focused on delivering primary storage, some that are focused on backup and disaster recovery (DR), some that are focused on archiving and some that are focused on collaboration.
So, let's consider [the] primary storage type of gateway, like StorSimple. In that case, caching, data deduplication, compression, snapshot technologies, cloning, cloud-based snapshots and cloud-based cloning. All of those are very, very crucial [for primary storage]. And I would add flash to that, too, by the way, because you're trying to minimize the latency.
That's the heaviest use case. That is the one that requires the largest number of features to be delivered. Now, if you look at DR, for example, and backup, then clearly, compression and [WAN] optimization is a common theme. But in a DR case, you'd want to have the appliance delivered either as a physical appliance or as a virtual appliance.
And why a virtual appliance? Because you want to be able to create that appliance inside the cloud, so that then you can either run the application in the cloud, so you have a whole infrastructure sort of recreated in the cloud to get away from this disaster; or you could have the application running in a remote location, but then you still need access to the storage, which is in the cloud. It can only be accessed via this cloud gateway appliance. So, having that appliance presented as a virtual appliance is crucial. So, that's a DR example.
We know that Microsoft just picked up StorSimple. Do you expect any more consolidation in the cloud gateway space?
Taneja: Let's look at why Microsoft bought StorSimple, and maybe that will give us some answers to this specific question. So, Microsoft wants as much data to be brought into Microsoft [Windows] Azure, their cloud, as possible. So, they effectively have bought this gateway to enable companies, enable their customers. How they'll package it and how they'll sell it, and whether they will make these appliances available for next to nothing to those that sign up for their services remains to be seen.
But the motivation for Microsoft is huge, because they have no legacy hardware to protect. So, it would be that sort of company most likely to look for another gateway product. Certainly Google comes to mind, and certainly there are other large cloud vendors that would have a motivation.
Now Amazon has gone and done their own [gateway], but the Amazon appliance is fairly weak. It is absolutely nothing more than a translator of iSCSI into REST [representational state transfer] protocol, or REST API [application programming interface]. So at some point in time, they might actually decide that they need a more powerful and a more complete engine or appliance. They might actually be a candidate as an acquiring company of some of these existing small companies.
How would a storage manager be able to identify a gateway product as weak -- or not good enough?
Taneja: So, the way to look at this, once again, is this way: What is this problem we're trying to solve? Are we trying to use cloud storage as primary storage? Are we trying to use cloud as a backup and for DR or archiving or collaboration? Once we start with that, then we look at the minimum functionality that needs to be built into this appliance to deliver a good result.
More cloud gateway appliance news
Cloud storage services aided by gateways
Using a cloud gateway appliance, object storage for archiving and backup
SMBs benefit from cloud gateways when using cloud storage
Cloud gateway-based primary storage becoming a viable option
And so, that's how I would start the process, and that's how I would determine if an appliance was complete in its architecture and its product incarnation. So, if you look at an appliance, and all it's doing is taking a flow of iSCSI traffic coming in from the data center side, and translating it into storage buckets, which are being used on the cloud side. That's the minimum level of functionality that you would want, but it doesn't have data deduplication, compression, WAN optimization. So effectively, if you got a hundred terabytes of data that is coming in from the data center side, from the application -- then about a hundred terabytes of data is going to go plunk itself on the cloud side, and you are going to then pay for 100 TB.
Well, what if I were to make that 100 TB go down to 10 TB? Well, I just saved a humongously large amount of money for that particular IT shop. Really, you have to, once again, start with what problem you are trying to solve.
What about file sharing and file sync? Is this something you see all gateway appliances having soon?
Taneja: That's a very good use case, by the way, for cloud. There are a vast number of products now that have become available from simple file sharing, like iCloud for Apple, for example. IT's very simplistic in the sense that I'm the user, these are my files, they happen to reside in the iCloud; and if I happen to have five, four, three devices, Apple devices, then those files will actually be made available in all of those devices so I can use an iPad, and iPod, et cetera, et cetera, and have a common [axis].
And then you have a whole series of other collaboration things that go all the way to an extreme level -- where five people across the world are working on the same file. And that requires very sophisticated locking, very sophisticated access methods.
At the appliance level, we do have companies like Panzura that are focused, first and foremost, on collaboration. So that gateway is, in fact, designed for collaboration. It might do a few other things, too, but the sweet part is collaboration.
So, I would see that over time, some of that functionality would get into some of the other existing gateway products, but collaboration done at a sophisticated level is a tough problem all unto itself. So, I expect that very specific appliances will continue to come to the market that are microscopically focused on collaboration.