Cloud Ideal For More Than Just A Data Repository

Processor

by George Crump

What if the entire storage management process could be rendered as simple as installing a cable modem? Simply connect a box to your network and then connect it to the Internet, and limitless storage that automatically manages itself is yours. This is the promise of cloud storage when used for primary data in conjunction with a cloud storage controller, a new breed of storage appliance that aims to make the experience of using the cloud much like using traditional onsite arrays but with cloud benefits.

Is this promise a reality for the SME? Can you really be using cloud storage for primary data?

Data Repository

Key Points • Cloud storage for primary data can reduce pressure on the backup and recovery process and automate much of the storage management process. • Cloud storage provides a built-in disaster recovery capability that can be ideal for the SME market that lacks the resources to equip and maintain a secondary site. • Be sure to find a cloud solution that’s specifically tuned for your applications or environment.

Cloud storage is often thought of as a repository for data that is not currently being used, such as backup and archive data. It is no wonder that the bulk of the initial success has been around solutions

that perform one of those two functions. Now, though, by using local storage as a cache and with some added intelligence, cloud storage is being considered for use as a primary storage area.

In fact, one advantage of using cloud as primary storage is that the need for separate backup and archive processes goes away. A file server that leverages the cloud essentially backs itself up via its replication processes, says Nasuni CEO Andres Rodriguez (www.nasuni.com). “Local storage is essentially turned into a cache to provide excellent local performance to the users,” he says. Rodriguez explains: As users store data on the cache, it is replicated to the cloud. Once in the cloud, it is snapshotted so that older versions of files can be retrieved in case of deletion or file overwrite. Snapshots of cloud data can be preserved as long as the storage administrator would like. As the local cache fills up, old data is removed. If a user accesses older data that is not in the cache, it is automatically recovered for them from the cloud.

The result is that several major interruptions to an IT administrator’s day are now eliminated. First, any data that can leverage this cloud architecture can be removed from the backup process. Second, the management of local capacity is now automated. No need to decide between deleting old files and expanding hard drives. Finally, user recovery of old data is now transparent. “Even files that have been deleted can now be recovered via a right-click within windows,” Rodriguez says.

Not Just File Services

Using cloud storage for primary data is not limited to file services. Ian Howells, chief marketing officer at StorSimple (www.storsimple.com), says, “A local storage appliance can be used that has enough capacity for active data and then leverage the cloud for less active data. This appliance could then present block storage via iSCSI that would allow certain applications like SharePoint and Exchange to use cloud storage. These applications in particular are ideal for a hybrid type of solution because they have a relatively small database that points to attached or related files. Those files quickly become inactive and are ideal for being stored only in the cloud.”

There are other advantages to an iSCSI gateway, says Nicos Vekiarides, CEO and co-founder of TwinStrata (www.twinstrata.com). “A key advantage of a gateway that provides iSCSI access is [that] you can use your own file system on that storage. For example, some administrators may want to use Windows for file serving, or others may want to use ZFS. iSCSI storage gives you the flexibility for both.”

Howells says the use of cloud storage on semistructured data such as email and collaboration relieves the IT administrator from tedious storage management tasks. All of the data is replicated to the cloud as it changes and then can be snapshotted for use later, he says, and the local storage automatically maintains the working set of data.

“There is no longer the need to add additional storage capacity or buy and install new storage systems. The cloud storage is automatically provisioned as the environment needs it. This also means the expense of costly system upgrades and the downtime they may bring is eliminated,” Howells says.

Not only does using cloud storage help resolve backup and storage management issues, it is also ideal for disaster recovery. Josh Goldstein, vice president of marketing and product management for Cirtas (www.cirtas.com), points out that many SMEs don’t have the luxury of a second data center for offsite backups or disaster recovery.

“In these situations, cloud-based primary storage is ideal. Data is inherently offsite and accessible from anywhere. Should a disaster occur, a standby cloud storage controller merely needs to be given access credentials to the data in the cloud, and all the company’s data for file services and key applications is immediately available,” Goldstein says. In a disaster recovery scenario, the cloud storage controller recovers data as it is being accessed, making the most effective use of Internet bandwidth.

One added benefit, Goldstein says, is that the failback to the primary data center after the disaster is easy. “Simply ‘remount’ the cloud with the cloud storage controller in the primary data center. No need to ship storage around the country or worry about copying all the data from a secondary site back to the primary,” he says.

Complement The Environment

Howells says the key for successful use of cloud storage is a system that complements the environment. “Look for cloud storage solutions that are specifically tuned for a particular application or use case,” he says. “It should understand how to integrate into that environment. Today excellent use cases are data sets where a large amount of the capacity is inactive but there is a need to keep it close at hand just in case it is needed again. Specific examples include file servers, email, and group collaboration software,” Howells says.

TwinStrata’s Vekiarides says that IT administrators need to consider the flexibility of the appliance to meet application performance requirements. “Different applications have varying needs and service-level requirements. Some applications may have different working set sizes. Some may be more latency-sensitive, and some may only be able to use the cloud for background replication with a full local copy. Translation? The caching intelligence needs to vary across applications as each has different attributes. At the same time, the caching has to be dynamic so that it can address changing application needs. A simple monolithic shared cache may not address these needs.”

When it comes to cost, Rodriguez says to make sure that the solution can provide space optimization capabilities such as deduplication and/or compression. “While you are paying as you go with cloud storage, you want to do everything you can to minimize costs over the long term,” he says.