When it comes to storing content for SharePoint, there’s a lot of confusion about what is supposed to be done. This is partly because Microsoft’s messaging about storage has shifted so many times that the answers are muddled, according to Chris McNulty, Quest Software evangelist for SharePoint.
As is too often that case, Microsoft has multiple ways to solve various problems. Users want straightforward answers: “Tell me how to do it, and I’ll do it. Don’t tell me I can do it this way or that way, or this way for now, but then I’ll want to move to that way later.” Microsoft needs to provide more leadership and fewer options. But, for an article on SharePoint storage, I digress.
Storage has become a big issue because more organizations are putting more things into their SharePoint installations. Microsoft’s guidance on how much data to put into the content database for SharePoint 2010 is 200GB. Meanwhile, organizations aren’t blinking about maintaining 1TB of data or more, and this huge amount of data impacts SharePoint performance negatively.
McNulty said the SharePoint team at Quest has identified five performance-killers caused by SharePoint content overload:
1 – The takeover of unstructured data. Unstructured data—Binary Large Objects, or BLOBS—contribute to nearly 95% of storage overhead in SharePoint content databases, according to Quest, and they take extra processing power and time to read the data.
2 – Avalanche of large media. Videos, images and other large media files are also occupying more space in SharePoint, meaning users will experience browser timeouts, slow Web server performance, and upload and recall failures, Quest has found. “It’s one thing if these are training videos that are actively used,” McNulty said. “But if that’s a secondary function, [those files] can be externalized.”
#!
3 – Old and unused files. Much of the content in SharePoint goes unused for long periods of time. Many organizations waste storage resources by keeping these old files around and backing them up as they do active content.
McNulty said this really is as much an ECM issue as a storage issue. “You need to understand the life cycle of information,” he said. “Create rules where content needs to be around for a certain length of time but doesn’t have to be in SharePoint. Have processes where things can move out of SharePoint into storage.”
An example he cited is work holiday lists. Workers really only want to search this year’s holiday list and don’t need prior years. That, he said, could either go into storage, or even be purged from the system altogether.
When SharePoint storage grows, McNulty pointed out, the biggest issue organizations run into is backup. “Big databases are inefficient to back up from SharePoint,” he explained. “Backup cycles can start to run four hours, eight hours, and that’s a practical pain point.”
4 – Not building to scale. As SharePoint content grows, Quest said, its supporting hardware can be underpowered. McNulty said there can always be a hardware solution to underperformance, but more efficient and less costly methods need to be identified. “Understand what SharePoint is intended to be used for, and this starts to give context around what you want in your SharePoint, and what you don’t want,” he said.
5 – Not leveraging Microsoft’s data externalization features. Microsoft’s first point solution for SharePoint storage was EBS: external BLOB storage, which was introduced for Microsoft Office SharePoint Server 2007 and is available and supported on SharePoint 2010. But there is no commitment to support beyond the 2010 platform; Microsoft’s committed, going-forward storage solution is now RBS: remote BLOB storage.
“Microsoft’s answer with RBS is to move data from the content database to somewhere else accessible with a provider interface,” McNulty said. Microsoft provides a free way of doing RBS with its file stream provider, which allows organizations to put content in file shares on the same machine as their SQL Server, he said.
So, taking the time to understand what you’re using your SharePoint installation for, and creating rules about managing the content, are the first steps to getting a grip on your storage problems.
— David
