ClearSky Data want to turn data storage into a utility service you rent from the cloud, but which is nevertheless available everywhere you are. It’s an ambitious goal.
ClearSky Data just announced a $27 million round of funding to help them pursue this goals, with content delivery network (CDN) provider Akamai one of the investors, and the round was led by Polaris Partners, an early backer of Akamai.
Essentially, ClearSky Data want you to rent storage life-cycle management from them. Instead of buying hardware, and managing the refresh cycle of upgrading storage controllers at the end of your depreciation period, ClearSky Data manage that for you for a monthly fee. You still use it on your site like a traditional SAN, but the actual storage becomes completely abstracted away.
It works by having a ClearSky Data supplied device sitting in your network, called an Edge Cache, which is an all-flash storage appliance. The Edge Cache could be on your own site, in your data-centre, or at a co-location facility. Your data is stored “in the cloud” and you connect to the Edge Cache device as if it were a physical SAN, and the fancy software figures out how to keep the most needed data close to you. ClearSky Data also have a cache tier in their metropolitan points-of-presence, which sit between the cloud proper and the Edge Cache.
ClearSky Data sees their initial customer base as being “solidly mid-enterprise”, big enough that they have enough storage to make operational management a pain, but not large enough to have really good processes in place to handle it. ClearSky is also consciously targeting CIOs who want to use cloud OpEx models, and are less focused on making a technical sale to the head of IT.
ClearSky Data have a partnership with DigitalRealty, a major co-location provider, and according to CTO and co-founder Laz Vekiarides “We’re going after the co-lo ecosystem in every metro first.” This is a smart plan, since co-location providers will already have excellent network connectivity, both with carriers and into the major public clouds that ClearSky Data uses. Customers at these locations are already half-way to having someone else take care of infrastructure for them.
The Akamai investment is interesting, because while there are some similarities to what ClearSky Data are doing and what a CDN does, you can’t really use ClearSky Data for sharing data between sites. The data is written back into the metro cache, but only gets pushed to the cloud every ten minutes or so (depending on the change rate) so the primary reason for having multiple sites would be for disaster recovery, not sharing data. That’s a pity, because sharing files between sites in a CDN-like setup is something of tremendous value to organisations doing CAD/CAM or media production who often need to share data files between offices. Others, such as Panzura, are attacking that part of the market, and ClearSky Data have signaled an intention to support NFS at some stage, so it might add the functionality later on.
Storing primary data in the cloud means you need reliable network access, and ClearSky Data has addressed this by partnering with network providers in metro areas, so access to the service is limited to where those networks have a point-of-presence, but where they do, it’s highly likely to be a robust and speedy service.