In this episode I’m with Joe Arnold (founder of SwiftStack) and we talk about:
– A short introduction to the company
– The differentiation between SwiftStack and others
– What we can soon expect from SwiftStack
Full transcript of the episode
Good morning, Joe. How are you?
Joe: Hi Enrico. It’s great to be on the program. I love the show.
Enrico: Thank you very much. You know, you know that the format of this show … This is just three questions in five minutes. I will go straight to the first question.
I’m very curious about Swift Stack as a company, and you as the founder. If you can introduce yourself and Swift Stack, that would be great.
Joe: I’d be happy to.
I’m Joe Arnold. I’m the Chief Product Officer, like you mentioned, of Swift Stack. Swift Stack is an object storage company. I got my start in object storage when I was at a company called Engine Yard. What Engine Yard did is we built a platform as a service for Ruby on Rails. We ran the Ru … Ruby … Ran a bunch of ruby interpreters in the rails project and Amazon came to us and said, “Hey, instead of doing your manage service of deploying on your own infrastructure, why don’t you use this new thing we’re building called EC2 to help deploy your customer’s applications?”
What we did was we built one of the early platforms as a service products on top of Amazon. In that environment, there really was only two kinds of storage. There was block storage, that people use to run data bases, run the virtual machines, and then there was object storage, and that was S3. What we did was we just changed how people were building and changed how they built applications so they could take advantage of S3 for the unstructured data that was part of their application.
I just got interested in deeper and deeper levels of the stack. When Open Stack came out, I was really early involvement in doing some of the early deployments of a competitor to S3, which came out of Rack Space, which was part of Open Stack, which was called Swift. We built that out for a couple of different service providers and realized, “Hey, there is a better way to build out an object’s … Build out Open Stack Swift, build out object storage, and put a product around that.”
That’s when we got involved with starting Swift Stack. What we’ve done here at Swift Stack is we’ve taken Open Stack Swift, we’ve built a product around it, which makes it really easy for enterprises to get something that’s like Amazon S3, but is behind their own firewall.
Enrico: That’s great. Actually, you know, I am quite a fan of object storage so … I’m really interested in what you are doing.
Joe: It’s just been really interesting because what’s happened is applications, now, have taken advantage of object storage in an incredible way. It expands from traditional enterprise applications like Back Up and Archive, like where you’ve done a lot of work with ComVault, with Veritas Net Backup, what’s now called Spectrum Protectively Storage Manager.
All these applications in the enterprise are converting over to use object storage because you can get so much more through-put out of that type of storage because it’s not a single stream application. We’re talking about these multi-threaded applications, if you will. Object storage is just perfect for that type of work-load.
Enrico: What is the difference between Swift Stack and other object storage products that we find in the market today?
Joe: The big differentiator is that we’re software that’s based on an open-sourced project. Not just an open-sourced project that, we’re the only contributors are, that we’re just carrying the flag. It’s an open-sourced project that’s powering some of the largest clouds on the planet. It’s powering Rack Space cloud files, IBM soft layer public storage cloud, it’s powering HPs, NTTs, and a whole ton of other very large clouds.
What we’ve done is we’ve taken that open-source core and we’ve made it very easy for deployment by enterprises. It’s sort of … Think of it this way, it’s been built and architect-ed for these large cloud environments, and what Swift-Stack does is we make it suitable for the enterprises. What that means is what we can do is we can, one, is they can deploy on a smaller footprint, they can pay as they go, they can scale up incrementally over time. We can start small for these enterprises, in the fifty to one-hundred terabyte range and we can have them grow up to multiple tens of petabytes, all within the same cluster on the same name-space, growing incrementally as they grow.
We’ve decoupled from the underlying hardware so that is can use the standard servers that they’re used to buying. They don’t need to buy an appliance from us, they don’t need to buy a particular skew from a systems vendor. We have certified skews with many different hardware vendors that are out there that they can use to add to their storage system.
Enrico: I’m sure that if we go on your website, which is SwiftStack.com, we can find all the [inaudible 00:05:51] on what is available today. Starting from that, what are you developing now? What can we expect for the next six months from Swift Stack?
Joe: I’m really looking forward to the future of Swift Stack. We’re doing some really great things around … We’re having a three-dot-oh launch that’s coming up. We’re announcing it at the Open Stack Summit. We’re announcing a few things. The first is that we have S3 compatibility. We have a number of application vendors that are supporting Swift Stack with the S3 API. It has a very high degree of compatibility with Amazon S3 and it performs great. We’re very happy about announcing that.
The second thing is we’re announcing erasure codes as a data protection strategy. Now we support both erasure codes and replicas for data protection and that works; That means that the application, depending on the need, can use either. We are … Then we’re making it even more …
The number three is we’re adding more capabilities around making object storage ready for enterprises. We have had a whole trove of integrations that work around authentication, utilization, and deployment. Now what we’ve done is we’ve made it much simpler to start small and scale up. We have all sorts of great tools that enable that, it enables Swift Stack to be deployed across multiple data centers. What that means is those three things together allows us to work with partners like Avere, who has a file caching [inaudible 00:07:53] appliance on the front end. Now people have a choice, they can put their data assets on premises, they can move those things to the cloud, it allows them to complete a great hybrid story around where they put data.
Enrico: Sounds great. We have a lot of reasons to stay in touch with you. I already mentioned that you’re website is SwiftStack.com, but how can we find you on Twitter or on our social networks?
Joe: Of course you can find the company at @SwiftStack, on Twitter. You can find me at @JoeArnold, or people can email me directly at Joe@swiftstack.com. Also, there’s a really cool thing for listeners to check out, I’ve written a couple of books on Swift. One is published by O’Reilly and another one is Object Storage for Genomics. People can go request a copy at SwiftStack.com/book, if they want to go check that out.
Enrico: Thank you very much, Joe.
Joe: Thanks, Enrico!