Wednesday, March 9, 2011

A cloud enabled Virtual Filer is software that runs as a virtual machine in a virtual infrastructure and doesn’t require any special hardware or storage. Instead, it leverages your existing virtual environment and uses on-premise disk to create a cache staging area for holding active data. As data is added or changes it’s replicated to a cloud storage provider of the administrator’s choosing, via intelligence built into the virtual file server software. When the local cache fills up older files are removed and stored only in the cloud storage provider’s system. The files still appear to be local though, because a placeholder is left behind in the cache. If a user accesses one of these place holders, data is transparently restored from the cloud service, often within seconds.

How Cloud Enabled Virtual Filer is Protected

How does cloud enabled Virtual Filer solve the biggest backup challenge facing data centers today? It eliminates the problem of big file servers. These file servers are often large both in terms of the capacity and the number of individual files they store. With a cloud enabled Virtual Filer the only data kept locally is that which is set aside for the cache, data which is also resident at the fully protected cloud storage provider.

This cloud enablement then creates the perfect protection strategy for file data, plus the Virtual Filer provides ongoing protection thanks to replication. As stated above when a new file is added or an existing file changes, that data is replicated to the cloud storage provider automatically on a regular basis, typically every hour. Most cloud storage providers will then replicate that data within their own infrastructure to multiple data centers.

Each change or version of these files is stored separately in a snapshot metaphor. That means that a user or administrator can browse through a hierarchical structure to find previous versions of files when a recovery is required. The versions of these files can also be kept indefinitely if needed.

Where cloud enabled Virtual Filers really shine is in Disaster Recovery. If the business needs to restart at another location in response to a disaster event, there is no significant amount of local data to move with it. There is also no file server hardware to ‘pick up’. At the remote site simply download the Virtual Filer software and bring it back online. Once the internet connection to the cloud is made the Filer and its data are instantly available to the users. No file recovery needs to happen since data is restored out of the cloud as users begin to access it. The result is instead of waiting for TBs of data to be transferred only MBs are actually recovered, and then, only as needed.

This is again ideal in the file server use case. With databases there are a finite number of files available and prioritizing those for recovery is relatively easy. File based data can number in the millions, if not billions of objects. Understanding which set of files should come back first could be impossible. The self-prioritization of the recovery is ideal since to the users it looks like all the data is back in place, when it’s only physically recovered as they access it. While the recovery of TBs of data from the cloud would be time consuming, recovery of MBs of data as needed is barely noticeable.

With these capabilities the Virtual Filer has to be considered “fully protected”. In fact a case could easily be made that this data is better protected than most file servers using standard backup applications.

First, protection is happening at least once per hour instead of once per night. This hourly protection is more appropriate for file based data, since most is extremely active from the point of creation to the point of finalization. In many cases changes to a file will happen all within one day, at most a few days. That means that a traditional backup may have only one copy of the file and often only the last copy. A cloud enabled NAS will capture an hourly version of the file and potentially eight or more copies of it. This gives much greater restore granularity if there is a need for a prior version.

Second, with traditional backup greater resource consumption occurs as the backup process churns through the millions of files on the servers. This churn requires file server processing cycles as each file is examined to see if it needs to be backed up. It requires network bandwidth as the identified files are transferred across the network and it requires backup server resources as those files are written to the backup device. Finally, it requires resources as those backups are copied and then moved offsite in case of a disaster. The resource requirement of the cloud enabled Virtual NAS is minor in comparison. When a file changes the file system flags it for replication. Then, when the replication time occurs just those files which have changed (typically, in the last hour) are replicated. Essentially, the backup and offsite data movement happen in one step.

As a business grows it needs to be able to share and collaborate on file based data, but the number of those files and their size is increasing rapidly. These files have a short, active life but a very high frequency of modification during that life. Both of these realities create problems for traditional backup applications. Cloud enabled Virtual Filers provide the file services that are required by the business and eliminate the problems that traditional backup software faces when protecting legacy file servers or NAS devices. To eliminate a big backup problem businesses are advised to look to the clouds.

George Crump, Senior Analyst