The size and capacity of a SharePoint Server 2010 implementation vary based on several factors, such as the number of concurrent users, service applications in the farm, the expected uptime service-level agreement (SLA), and others. These factors dictate how many servers are required in the SharePoint farm and how the overall farm architecture looks. Also, when looking at the infrastructure design decisions for multi-tiered applications such as SharePoint, virtualization technologies should be considered to achieve higher utilization levels and reduce the overall infrastructure costs. Virtualization results in additional benefits, such as reduction in infrastructure costs required to deploy the application farms and enable dynamic on-demand scalability of the application.
Virtualizing a SharePoint 2010 application deployment requires certain best practices be followed to achieve optimal application performance, to ensure high availability of the farm components, and to realize the additional benefits mentioned above. The three-tier architecture governs the best practices to be followed and requirements while deploying SharePoint Server 2010 in a virtualized environment. This includes factors such as resource (CPU/memory/storage) sizing for the present needs, scalability to support to business growth, and planning for high availability of the application infrastructure.
However, SharePoint Server is a platform for achieving various things. No two customers use SharePoint the same way. So, the sizing aspects become little tricky compared to other application workloads. This is where we use a combination of factors such as workload usage profile, number of concurrent users, and number of requests per second desired in the implementation to determine the right size of the hardware to support the SharePoint workload. This sizing guidance should include enough room to support growth in the near future, at least. The other aspects, such as high availability, are typically the characteristics of the underlying infrastructure. For example, does the infrastructure used for virtualizing an application like SharePoint provide support for easier scalability? Does this infrastructure support failover and high availability features for the applications? And, if so, how does it complement the application level high availability features?
Once we have the right sizing and other design aspects figured out, we need to implement the application architecture and validate the same against the requirements we laid down before the design phase. This is the most important step in the process to ensure the infrastructure and application architecture supports the minimum requirements and meets the expectations.
As a part of Dell’s Enterprise Solutions Group (ESG), we published a detailed study of designing and implementing SharePoint Server 2010 on Active Infrastructure. In this design and implementation guidance, we addressed the steps mentioned above and provided guidance on implementing a virtualized SharePoint Server 2010 farm for 5000 concurrent users in a SharePoint collaboration usage profile.
As an extension to the design and reference implementation guide, a performance study was conducted to understand the overall reference implementation capabilities and how the Active System 800v infrastructure design principles help the application enable high availability. This performance study was performed using Dell SharePoint Load Generation framework.
For detailed best practices, configuration guidelines, performance validation procedures, and results, please refer to Microsoft SharePoint Server 2010: Design and Implementation on Active System 800v.