Sanity for SharePoint Storage

Editor’s note: Below is a gust post from Howard Marks. He is the Chief Scientist at DeepStorage, LLC. an independent test lab and analyst firm following data storage and information management. You can read more about him here. Here’s Howard’s post:

————

It should be a surprise to no one that SharePoint has been the most successful server product in Microsoft’s history. By the time SharePoint came to market, corporate America was desperate for a better content management system than email and unmanaged file shares.  Recent surveys have shown 74% or all organizations and 80% of the Fortune 500 have at least pilot SharePoint implementations.

While SharePoint is a big step up from local file shares and sending files around the organization as email attachments, many organizations are struggling with managing their SharePoint infrastructures as  SharePoint and database servers proliferate to keep up with the flood of data.

From a storage management point of view SharePoint’s biggest problem is that it stores all of the documents and other files that it’s managing as SQL Server BLOBs.  Storing documents as part of the SQL Server database, along with indices and other metadata, means a typical SharePoint server farm needs ever expanding amounts of expensive high speed storage.

True, SharePoint administrators can export SharePoint documents using SQLServer 2008’s Remote Blog Storage and FILESTREAM data type but this feature can only export BLOBs to a local file system on the SQL server.  Once the BLOB data is exported from the database, administrators have to carefully coordinate backup jobs to ensure that the exported file system data and the SQL Server metadata that describes it are backed up coherently.

Dell has taken on these data management nightmares with a SharePoint Infrastructure Optimization solution that leverages the DX object storage system and Ocarina compression technology. With the Dell solution SharePoint administrators can take control of their SQL Server storage by doing much more than simply migrating blobs to DX objects.  The AvePoint DocAve SharePoint management suite at the heart of the solution allows administrators to build policies that migrate data from BLOLBs to the DX  based on SQL Server storage consumption, document age, and/or document access history.

As we’ve already seen simply migrating blobs to an external data store is at best a partial solution. The Dell SharePoint solution not only optimizes the SQL Server backend by exporting blobs but goes a step further to SharePoint aware information lifecycle management. Organizations typically will build SharePoint site collections around some project. In a classic SharePoint environment that data stays in the production SQL Server long after the project is completed. Even if the blobs have been exported to some other data store the entire site collection has to be managed and backed up with the same level of protection as site collections related to much more active projects. The Dell solution allows administrators to migrate entire site collections Windows projects are completed.

As data is migrated to the DX its compressed using Dell’s Ocarina file type specific compression which in this type of application should be substantially more effective than even the combination of generic compression and data deduplication. Migrating data from the 15 K RPM drives or hybrid storage that supports SQL Server to a cost-effective, almost infinitely scalable, object storage system like the DX will not only reduce storage costs directly but also reduce the number of SQL Servers required to support a SharePoint farm.

Dell solution also addresses the nightmare that can be SharePoint backup, performing a holistic backup of the entire SharePoint infrastructure not just the SQL Server database and document data but also configuration information like IIS settings and the IIS metabase. Of course objects that have been migrated to the DX platform are safely archives so they needn’t be included in multiple generations of backups as they would be with a more conventional system, again reducing storage costs while speeding backups and more importantly restores. Once data has been migrated to a DX it is replicated across multiple nodes in the DX cluster and between multiple DX clusters, hopefully in multiple locations.

All in all Dell SharePoint infrastructure optimization solution looks like it could bring some substantial amount of sanity to SharePoint storage management and make SharePoint administrators lives substantially easier. As a longtime system administrator myself anything that makes administrators lives easier is a good thing for me.

About the Author: Howard Marks