Load balanced BI Solution - Storage and deployment strategy

In a MS BI platform based solution, Sharepoint it mostly used as the deployment platform for web-applications, reports and other artifacts of the solution and it's a very standard practice. But this remains as simple as it sounds only till you have one server per technology / hosting platform, for eg Application Server, Database Server, Report Server, Analytics Server, ETL Server etc.

As soon as you start scaling your solution infrastructure, like load balanced federated server solution, implicitly your solution gets replicated across servers. There are standard practices to keep the codebase synchronized, but the caution area that is not visible implicitly is that Sharepoint itself can act as a data storage medium for a variety of content. In fact Sharepoint is recognized as a Content Management System. You can store SSRS reports, Performancepoint artifacts, Excel Services worksheets, Visio diagrams and much more on sharepoint. If you are not well informed of the internals of how sharepoint stores data, you would end up replicating tons of these artifacts on different servers.

The key piece to this issue is a one piece of information - every sharepoint site has a content database and that database is a SQL Server based database. Application can store data in sharepoint in sharepoints lists and other objects, and BI solution artifacts are also stored on sharepoint sites. When you replicate your sites to different servers in a load balanced environment, all sites should refer to a common content database, and configuration should be done accordingly for all the sites.

This strategy gives a huge benefit, the prime ones being centralized deployment and bullet proof shield against data integrity issues. Anything stored on one site is automatically available on all other sites, avoiding any synchronization between sites replicated on different servers. Also you deploy artifacts on one site, it would get stored in shared content database and effectively the same becomes available on all the replicated sites.

Dimensional Modeling, Cube Design, Dashboard Design, ETL Design, Reports Design etc are popular BI areas, but solution design is the ultimate expectation from an Architect. Professionals can shy away by taking refuge under their strength area like Data Architect, Application Architect, BI Architect, Integration Architect etc, but someone would have to assume role of driving these genius and this class of architect is known as Solution Architect.