GigaOM Pro

Source: Matthew Field

Cloud storage got a big shot in the arm this week in the form of tens of millions in venture capital and research funding. Why? Because despite all the talk about applications and servers in the cloud, the real value lies in data, which must be housed somewhere.

Why Cloud Storage Matters

Cloud storage represents a major opportunity for organizations because it lets them store terabytes of data without paying to build and maintain traditional systems. For new web-based startups, cloud computing might fill the entirety of their storage needs — primary, backup, archiving, whatever. For more traditional businesses, the value lies in backing up primary data and archiving less valuable data for compliance reasons. Archiving might seem a low-value job compared with primary storage that serves applications, but, as we’re seeing with the advent of Big Data tools, there’s value waiting to be extracted from everything, even long-forgotten files and documents.

Even for organizations that choose to maintain their own storage systems, cloud technologies have made the ability to scale easier and cheaper. Companies no longer have to buy over-provisioned systems from traditional storage vendors. On the file system front, in particular, cloud software lets organizations build scale-out systems on commodity servers, adding volume as needed at the level needed. In some cases, as with Caringo’s CAStor offering, this requires little more than plugging in an additional box.

One recent calculation has enterprise storage costing $25 per gigabyte per month and consuming 40 percent of IT budgets. Although not an apples to apples comparison, pricing for storage with cloud storage provider Nirvanix begins at 25 cents per gigabyte per month.

It’s no wonder a 451 Group report released this week estimates that cloud storage will account for 40 percent of $964 million in “core” cloud revenue this year. For an overview of some of the more-promising startups, including Nirvanix, Zetta and Gluster, which combined for $30 million in funding this week, see my post on the Structure blog from earlier this week

What’s Being Done

Cloud storage is not perfect, however; there is work to be done in areas like security, lock-in and reliability, in particular. The latter two issues were addressed in part this week. IBM Research is tackling interoperability thanks to a $21.5 million (€15.7 million) grant from the European Union. Data portability might be the biggest concern around lock-in, because of number of proprietary storage formats among cloud providers. If customers want to move data from one cloud to another, they often have to download their cloud-stored data, convert it into a new format, and then upload to the new provider.

The VISION Cloud project, led by IBM Research but comprised of representatives from various industries (software, service provider, telco, standards bodies and research), aims to improve cloud storage by developing capabilities across four fronts, including “full data interoperability.” According to IBM Research’s Yaron Wolfsthal, the project will be complete in three years, at which time it will release an open-standards-based framework. Cloud standards efforts haven’t been overly successful to this point, but this project will produce some, which is a good start. The other three targets of VISION Cloud — rich object data models, computational storage and content-centric access — will improve the performance and searchability cloud storage platforms.

Targeting reliability, particularly of the Internet, Amazon Web Services released its Multipart Upload feature for its S3 storage service. Among its other reasons to exist, Multipart Upload ensures that users need not re-upload large objects in their entirety should network issues stop the process in midstream. By using the feature, users can break objects (up to 5GB) into smaller parts (5MB or larger) that upload individually. When all the pieces have been uploaded, the object is finalized as whole within S3.

It’s true, of course, that cloud storage has a long way to go before it approaches the performance and features of on-premise systems, but the promise is evident. Customers are buying it and VCs are funding it because storage is important, and backing up data in the cloud is cheap, easy and increasingly risk-free. And because for a growing number of small companies, it’s the only way to go.

Question of the week