To say SharePoint 2013 is resource-hungry might be putting it lightly. In his sessions for STP 2013, Michael Noel delved into the power infrastructure needed to run the latest in collaboration with Microsoft.
One of the reasons you need more power is that the new version of SharePoint does so much more on the back end. For example, there is now an integrated translation service that sends documents up to the cloud through Bing translation services, translates the documents, and then sends them back to your desired destination. Staying on your farm, there are new Work and Request Management Services that keep your servers running equally and can even direct incoming traffic to specific servers based on where it’s coming from (iPad, Firefox, etc.). SharePoint 2013 needs more power but also provides the leash necessary to keep it all under control.
Other major changes Michael talked about are profile syncing and authentication. The new User Profile Sync gives you three options for the deployment of varying complexity. The simplest, AD Import, is one-way sync much like SharePoint 2007; the middle ground is the two-way sync you’re used to from SharePoint 2010, and the most complex is Forefront Identity Manager (FIM) for use in large organizations. As for the authentication, the default is now claims-based authentication. While SharePoint 2013 does offer support for classic authentication, Microsoft has said that classic authentication is going to go away and claims-based will be the norm. Michael highly recommends updating all authentications to claims-based before upgrading to SharePoint 2013.
Data management, server space, and power are the keys to a successful SharePoint deployment. To keep all this exciting new stuff running, you need to master all three. Michael breaks down server setups and tells us what Microsoft is doing to help with data management. As you might think, the more servers, the better, but using them efficiently is just as important. In his slide deck, Michael has several tables showing minimum suggested resources as well as several server configurations.
The basic takeaway is to protect yourself with an “always-on” system. These systems often consist of a mix of physical servers and virtual servers (slides 13-18 in his deck show different setups). Being able to mirror a server will enable you to ensure your SharePoint environment stays up and does not lose any data. Michael’s suggested method to achieve this is through SQL 2012 AlwaysOn Availability Groups (AOAGs). Speaking of data, Microsoft has redesigned data storage for SharePoint 2013. Shredded storage will be your new best friend.
In the past, SharePoint would store any changes to a file in a new version of that file while keeping the original, quickly creating large BLOBs. Now, with shredded storage, SharePoint stores the original file and then smaller delta files associated with the original. In the old system, one megate change to a 50 megate file would result in a 51 megate file AS WELL AS the original 50 megate file. Now with shredded storage, that one megate change will be stored as a one megate file leaving you with the same data but using only 51 megates of space total instead of the 101 megates of the old way. We’re talking about serious space-saving here.
Other best practices Michael touched on are keeping databases focused and mirroring everything (AOAGs). With focused databases, you save space and improve performance. Using the AOAGs allows a safety net in case something happens and one server goes down. Lastly, if you are just hopelessly lost in an endless pit of data, or if you’re a standard end-user, Microsoft has added FAST Search. What used to be an expensive add-on is now included with SharePoint 2013 (not the free version though). This allows for data-rich searches of entire farms, making it fast and easy to find that one file that you know you had but just can’t quite put your finger on.
