Feature

The cloud washing game: Key functions of a private storage cloud

Ubiquitous use of the term cloud storage has brought more confusion than enlightenment to the industry, leaving data storage administrators struggling to avoid cloud washing and trying to figure out what constitutes a real cloud. When it comes to private storage clouds, it’s often unclear where the cloud begins and traditional centralized network storage ends, muddling the picture even more.

If you aren’t sure of what a true storage cloud is, you’re not alone. Even organizations that have deployed one may not know the definition of a true cloud.

“Some customers either love or hate the word 'cloud,'” said James Sangster, senior director of solutions marketing at NetApp Inc. “Some will say ‘I don’t have a cloud,’ but by the strictest definition, they do have a cloud. They just don’t call it a cloud. Others say they have a cloud, but at a closer look it’s just a bunch of centralized storage. It’s what we’ve had for decades.”

Location is the first part of understanding a private storage cloud. A private storage cloud can operate in a co-location but it typically stands on the company’s premises. Many data storage administrators want the benefits of a cloud while they retain control over their corporate data. “You create a public cloud construct and the customer says ‘I love that but I want that in my four walls,” said Steve Zivanic, vice president of marketing at Nirvanix Inc., which provides the technology infrastructure and functions that enable cloud storage for companies. “All the customer has to provide is a data center, and the power and cooling. We do everything else. It’s fully managed as a service.”

When having cloud conversations, it’s helpful to think of the cloud as a delivery service and virtualization as a technology to facilitate cloud implementations. There's also a list of key functions that separate private cloud storage from a traditional storage configuration, according to many experts.

Object stores: Object–based storage is a fundamental building block of a storage cloud because a true cloud offering must handle billions of objects translating into hundreds of petabytes of data. Object-based storage is the best technology for growing storage linearly and making it elastic.

“I think object storage is the future,” said David Vellante, founder and chief executive officer (CEO) at analyst firm Wikibon. “A true cloud scales to petabytes, billions of objects. A traditional disk drive [or file system] can’t do that.”

Not all applications support object stores yet, although the number of those that do is increasing rapidly. “More applications are supporting object stores, especially backup applications like CommVault Simpana and Microsoft SharePoint,” said Jerome Lecat, CEO at Scality Inc. “A few years ago, there were very few.”

Linearly scalable: A private storage cloud has to be linearly scalable, meaning there's no limit to the amount of capacity a company or user can access. “This is where it gets obfuscated,” said Marc Staimer, president at Dragon Slayer Consulting. “Traditional storage vendors have a storage limit; they have hardware limitations.”

Still, many of those same vendors are rebranding their products as belonging in the cloud category -- despite their limitations -- and Staimer is among those who say that this sort of rebranding is an example of cloud washing.

On-demand resources: Resources such as capacity are supplied on-demand in a private storage cloud. The cloud makes it easy for users to get more storage capacity automatically, eliminating the need to contact an IT administrator to allocate it manually. “It’s the ability to order compute, storage or networking capacity dynamically by just saying ‘I need more,’” Wikibon's Vellante said. “Capacity is there, it just needs to be turned on. Before, when you needed capacity, IT would call up EMC and say ‘Send some boxes over.’”

Geographically aware: Private cloud storage is geographically aware, meaning the system is smart enough to know where the data resides so it can be accessed efficiently. This sort of geographically aware system is a function of object storage. The system knows where the data is and where the user is and, based on changes in patterns, will replicate or write data where it's most likely to be accessed. “It understands locality and says ‘This is the closet area you can get your data,’” said Tony Asaro, CEO and founder at INI Group.

Chargeback and multi-tenancy: These two aren’t mandatory for a private cloud, but can be valuable in many cases. In traditional storage environments, a chargeback system is common. In that model, each department or division’s storage use and transactions are tracked so IT administrators can see where the bulk of the storage burden is.

NetApp’s Sangster said many of his company’s cloud customers aren't using chargeback yet. “A lot [of them] are doing ‘showback,’ where they show the department what it uses, but the company isn't going to charge,” he said. The difference between “chargeback” and “showback” is the former places a dollar figure on the amount of storage resources used, while the latter places a percentage figure on the amount of resources used.

Multi-tenancy is required when customers share an application or resource, and each customer’s resources and data can't be available to other tenants on the cloud. While this is a mandatory function of public clouds, it’s not always necessary for a private storage cloud. “Not everyone with a private cloud cares about this,” Dragon Slayer Consulting's Staimer said. Some will use multi-tenancy to restrict data usage among departments, while others won’t deem it necessary.

Tech refreshes and data migrations: Two key differentiators between a private storage cloud and a traditional storage implementation is that cloud customers no longer have to deal with technology refreshes and data migrations. This is a tremendous benefit for storage administrators because migrating from an older system to a new one is often an onerous task -- one now handled by cloud storage vendors.

“You have eliminated that issue [with the cloud],” Staimer said. “A tech refresh in cloud storage is a non-event. A new node is added and data will automatically move. Think of it as a tank of water. You can add another tank and it will find its own level. You never have to reboot a storage system and you never have to redirect servers to the storage.”

Nirvanx’s Zivanic summed up the difference between a storage cloud and a traditional storage deployment this way: “If you pay maintenance fees or do a tech refresh, that isn't the cloud.”

BIO: Sonia R. Lelii is the senior news writer for the Storage Media Group.

This article was previously published on SearchCloudStorage.com.


This was first published in October 2011