The Shortcut Guide to Storage Considerations for Microsoft SharePoint

by Wendy Henry


Though Microsoft SharePoint has emerged as a market front-runner in the race to empower business users and provide collaborative systems for decision making, few SharePoint installations live up to the prophecies originally hyped during the buy-in process. The brilliance of any successful enterprise application is its sustainability, not the initial reply to an urgent business need.

The Shortcut Guide to Storage Considerations for Microsoft SharePoint
will help bright IT engineers and administrators choose flexible storage for SharePoint that will stand the test of time. The guide examines SharePoint’s distinct storage requirements and recommends strategies for determining optimal yet affordable disk solutions that will support your SharePoint investment both now and into the future. Readers will learn about the pitfalls of inadequate storage management, I/O performance enhancements for SharePoint components, and high-availability options combined with data archiving to best manage data bloat. Lastly, an in-depth discussion of best practices for deploying SharePoint onto iSCSI will prepare any enterprise, large or small, which may be ready to ramp up their disk architecture.


Chapter 1: SharePoint Capacity Requirements

In March 2008, Bill Gates revealed that SharePoint had joined the ranks of Exchange and SQL Server by becoming one of Microsoft’s fastest growing network applications, surpassing 100 million licenses sold and anticipating 2008 sales in excess of one billion US dollars. The seed idea of using Microsoft’s collaboration platform as a viable alternative to traditional file storage systems has germinated among enterprise network administrators thanks to SharePoint’s improved SQL Server repository and mature document management features added in 2007. Integrated content management capabilities have caught the fancy of Web masters looking for a simple, flexible alternative to heavy ASP.NET development on traditional IIS hosting. In fact, some enterprise network implementations of SharePoint include such diverse custom programming that the application has become a platform unto itself, supporting not only data delivery but client software interface delivery as well.

With so many reasons to prefer SharePoint as an all-in-one solution for disseminating information, it stands to reason that an integral factor in any solid SharePoint deployment plan should be storage. And yet, selecting an appropriate storage system solution is often overshadowed by elaborate governance and logical design planning. Perhaps many procurement engineers simply rely on the old “more free space is always better” adage by purchasing big hard drives in the server hardware. Or maybe the SharePoint planning committee is too overwhelmed by the unknown aspect of user content contributions to accurately choose an appropriate storage system for SharePoint at the onset of deployment. Whatever the reason, SharePoint suffers one of the highest content migration rates within the same version when compared with other Microsoft enterprise server applications.

Chapter 2: SharePoint Storage Architecture

Based on the information explored in Chapter 1, it should be apparent that analyzing SharePoint storage demands more than just a single focus. Because there are so many Microsoft applications and services at work, not to mention the Windows OS itself, it is important to take each individual product into account when determining the best storage solution. From the use of Virtual Memory Manager in Windows to the art of optimizing SQL Server 2005 databases, there are many considerations to examine. This chapter will dissect the OS and each application supporting a SharePoint installation—in particular, the many SQL Server 2005 databases required for SharePoint will figure prominently as well as IIS optimization and SharePoint indexing behavior.

When it comes to managing storage on any Windows Server OS, there is more than meets the eye. And although it may seem simple to concentrate solely on the I/O subsystem of hardware and the disk controller driver software, doing so would only expose half of the picture. In Microsoft, as with many platforms, storage management has an additional influence that can affect readings and decrease disk efficiency. This influence is memory management or, more specifically, the Memory Manager. The first section of this chapter will expose the Memory Manager and its effect on storage performance.

Chapter 3: Best Practices for Deploying SharePoint on iSCSI

If the previous chapter emphasized nothing else, it stressed the benefits of employing a fast and vast disk architecture for the components of SharePoint. Everything from the OS through IIS to SQL Server make good use of multiple disks, and the scaling of SharePoint across multiple servers demands data transfer via speedy, reliable protocols. Settling on a storage solution for SharePoint and its burgeoning SQL Server databases often has less to do with budget and more to do with long‐term investment. Those who underestimate storage needs and purchase minimal storage in an effort to reduce the initial cost of implementing SharePoint invariably pay for their decision in the long run when they have to replace insufficient systems that cannot be expanded.

Of the three storage architectures described in the previous chapter (Internal/DAS, NAS, or SAN), the most popular, and perhaps appropriate, strategy for a SharePoint enterprise is SAN. Though by comparison the SAN is the most complex storage strategy to install and manage, its implementation and administration costs are quickly overshadowed by its extensive availability and recovery options, immense scalability, and autonomous configuration flexibility. But saying you will choose a SAN over Internal/DAS or NAS is like saying you are buying an automobile instead of a skateboard or bicycle…the question becomes: what kind of automobile?

There are different SAN solutions, each employing their own transport protocols and medium architecture. In this chapter, we will focus on the Internet Small Computer Systems Interface (iSCSI) mass storage/networking protocol that operates over various mediums. The iSCSI protocol is an affordable, popular SAN implementation that lends itself well to Microsoft Windows networks that rely on TCP/IP for a transport protocol. Offering a wide array of disaster recovery and data availability solutions, most iSCSI SAN providers capitalize on both the reliable storage and speedy delivery data needs of SharePoint.

Chapter 4: High Availability, Disaster Recovery, and Maintenance

Regardless of how fast or efficient your storage solution for SharePoint is, the entire SharePoint infrastructure is vulnerable if you do not engage in regular maintenance and employ some type of data protection strategy. SharePoint is an organic collaboration system that will likely host volatile mission-critical data. Sustaining data availability and delivery speeds the SharePoint users have come to expect should be a top priority for any SharePoint administrator. SharePoint usage will evolve over time, so any high-availability (HA) or disaster recovery (DR) plan should be routinely reviewed, tested, and potentially modified to accommodate the system.

The terms high availability and disaster recovery have become almost synonymous in today’s IT industry. Understandably, the need to provide persistent and quick access to critical data almost certainly entails the need to reproduce that same critical data should it go missing. But in truth these terms express two different phases of an overall goal. HA SharePoint solutions are strategies that provide redundancy and workload balancing to deliver data quickly to users without detrimental interruption. Depending on the HA solution employed, a noticeable interruption of data delivery can be virtually eliminated or at least mitigated by setting reasonable expectations. HA solution performance is often measured by the uptime of the system (the percentage of a time slice that the system is online and delivering to the users). For example, the ultimate goal of an IT system may be 99.999% uptime or less than 6 minutes downtime per year.

DR strategies, however, define methods of restoring critical data that has been rendered extinct to the SharePoint infrastructure. Think natural disaster, irreparable disk failure, or irreversible data corruption. And though most DR strategies are built around a redeemable backup operation for valid reasons, leaning on an alternative HA redundant data source can also provide some measure of recoverability. In fact, many of the new iSCSI SAN data bit replication offerings in the market are making a name as DR solutions as well as HA stratagems. Choosing an appropriate DR scheme will depend on your Service Level Agreement (SLA) and user demands.

Facilitating HA and DR strategies will require monitoring as discussed in previous chapters and routine maintenance tasks. Data may occasionally need to be relocated via volume migration to support online, near-line, and offline availability. Load balancing may involve automating the rollover of workloads in response to dynamic user activity on the SharePoint system. This chapter will discuss hardware and software availability challenges, common DR strategies, and routine maintenance operations that enhance a SharePoint environment stored on a SAN infrastructure.