On Monday, SwiftStack closed a $16 million Series B funding round to help the startup bolster sales and marketing...
of its OpenStack Swift-based object storage software.
The San Francisco-based vendor sells a commercially supported version of open source OpenStack Object Storage, more commonly known by its project code name, Swift. SwiftStack tacks on a proprietary management system and a file system gateway that extends the product's functionality beyond the scope of OpenStack Swift.
SwiftStack was incorporated in November 2011, and secured $1.5 million in seed funding the following May, followed by $6.1 million in Series A funding in March 2013. OpenView Venture Partners provided a substantial investment for the current funding round, joining existing investors Mayfield Fund, Storm Ventures and UMC Capital.
OpenView Venture Partners specializes in scaling out enterprise business software companies, according to SwiftStack. Operating Partner Daniel Demmer will join SwiftStack's board. He was formerly president and CFO at Endeca Technologies, which Oracle acquired in 2011.
On the eve of its three-year anniversary and the upcoming OpenStack Summit in Paris in November, SwiftStack CEO Joe Arnold discussed the company's progress, competitive landscape and future direction.
In part one of a two-part interview, Arnold said SwiftStack's customer count has grown to 40, with storage capacities ranging from 100 TB to tens of petabytes. Customers include eBay, Hewlett-Packard, Pac-12 Networks and Time Warner Cable.
What are the most prominent customer trends you've noted during the last year?
Joe Arnold: What has surprised us the most is that these large-scale enterprise companies that are running big data centers want to start consuming open source in storage. That was the biggest eye opener. When we first started the company, we were focused on people who were doing Web-scale work, and that was great. They were already there. It was preaching to the choir. They were used to using open source software and pairing that up with commodity hardware components. We were pretty focused on that. And what we noticed is that we kept getting called into these larger enterprise accounts, which would normally be the foothold of some of the larger storage vendors. And they were trying to wrap their arms around it. More of them are starting to click and say, 'OK, we can take on software-defined storage, and we're going to use it for new applications. And we're going to be using it for backup and archive workloads.' They're inching their way into this [technology] occupying more and more of their data center.
Do you find their use is restricted to backup and archive? Or are they thinking about extending the approach to primary storage?
Arnold: We absolutely get the backup and archive, and document management software can layer onto what we do. Video and media assets can get layered on. What we're not doing is running databases and virtual machines. They're still running databases and virtual machines on their traditional infrastructure. But we're coming alongside and offloading that type of data, which actually is better served with an object storage solution because you don't need to do things like file locking or you don't need to take on synchronous workloads. You can push documents to it. You can push backups to it. And it allows them to not have to deploy quite as much Tier-1 storage.
Are customers typically trying to shift from tape to disk-based backup and archive?
Arnold: Usually, it's with people who are on a filer, and they want to move to object. They want to do that because setting up multisite for disaster recovery purposes usually requires more expensive software or more infrastructure because of that setup. They'll use the same backup software, and then instead of pointing it to a filer, they'll point it to the object storage system, or they'll point it to SwiftStack in our case, which behind the scenes is just putting data in multiple sites.
Is the ability to use commodity hardware the main advantage for them?
Arnold: Cost does play a big factor. For how we get leads, that's probably No. 1. And then No. 2, they're looking to do multi-data center. There are other products that will do the multisite, with hardware products or some software products. You think [EMC] Data Domain. They can go in and not use those products and instead use the multisite capability.
Do customers have to change their backup software?
Arnold: With people who are running say a big virtual machine infrastructure, even with VMware, they'll use backup software to take the snapshot. … Then there are other people -- eBay's a good example -- where they have databases they're backing up. And they have their own tools to create those database backups, and then they'll push them into our system. The reason why SwiftStack can do that is because we've introduced the file system gateway product. So, we don't necessarily need a replacement of the backup software they're using. We can just use the existing backup software they have in place, and that file system gateway is a perfect target for accepting those backup files from the backup applications.
Do you see a spillover effect from other OpenStack projects?
Arnold: Yeah. We like to be another tip of the sword, if you will, of getting OpenStack in more places. Not everywhere is going to replace their compute environment, because that's a big project. But it's really easy for them to take on OpenStack Swift and SwiftStack to offload some storage use cases they may already have. So, we can enter into environments more easily and I think start getting a foot in the door for the rest of OpenStack.
How do you plan to use the $16 million?
Arnold: We're of course going to continue to grow our product development team, but we haven't really spent in a large way on sales and marketing. So, that's where we're going to start spending money.
Using open source SDS for primary storage projects
OpenStack Summit reflects momentum for Swift object storage
Storage is key piece of overall OpenStack cloud picture