Monday, 21 May 2012

Three-Tier SharePoint 2010 Architecture

Uninterrupted SharePoint 2010 Farm
-  Scalable SharePoint Farm 
-  wfe cluster scalable upto 32 nodes
-  High availability SharePoint 2010 Farm
-  SharePoint 2010 Fail-over Farm

I had a great time during the migration process from MOSS 2007 (2-Tier) to SharePoint 2010 (3-Tier) architectural transition.
..migration was completed in April 2012.

Three-Tier SharePoint 2010 Architecture
Three-Tier SharePoint 2010 Architecture
network diagram showing 3-Tier SharePoint 2010 Farm: web front-end layer, application layer and database layer. The above architecture was successfully designed, tested and implemented by me.
The above scenario is scalable and highly available systems reliably provides an acceptable level of service with minimal downtime and finally easy to manage & maintain during patch management schedules.

Our aim was to:
* verify that each web request (http port 80) were distributed automatically to the available cluster WFE servers. .yes that’s how it works..!
* Shutdown WFE02 and make sure there is minimal downtime and automatic load distribution.. yes in case of abrupt failure of 1 web front end node, clients notices fraction of seconds delay in web response time however redundancy works magnificently as expected..! we will have sufficient time to plan and rebuild the failed server in such scenarios.
* Shutdown application layer and see if the intranet works uninterrupted... yes intranet works still uninterrupted..! coz.. 'Microsoft SharePoint Foundation Web Application' service runs only on WFE servers. Two other services which we enabled on wfe servers are: i) 'Search Query and Site Settings Service ' and ii) 'SharePoint Server Search' services.
* Shutdown Active SQL server and see if the Passive SQL server takes over or not.. yes it works..!
* Performance was superb..!

Three-Tier SharePoint 2010 Architecture 
showing uninterrupted farm in the event of multiple failures
Three-Tier SharePoint 2010 Architecture

network diagram showing 3-Tier SharePoint 2010 Farm with fail-over scenarios

microsoft reference

I have explained in my blog quick clustering steps for SQL 2008 R2 cluster and fail-over farm including best practices.

next steps:
our next step is to setup disaster recovery site and setup a test and development intranet environment.
so we have to scrap the existing disaster recovery intranet site and upgrade the architecture from 32bit to 64 bit OS.. from Server 2003 to Server 2008 R2 Enterprise and SQL 2005 32bit to SQL 2008 R2 64bit Enterprise edition environment and then upgrade Microsoft Office SharePoint Server (MOSS 2007) to SharePoint 2010. And finally start real-time replication from production farm to the disaster recovery farm.

currently we are using DocAve from AvePoint for real-time replication of SharePoint portal.
we can also achieve this directly from SQL DB replication technology, either via log-shipping or mirroring. however we will test this as the current databases are huge. ..I will shortly scribble the procedure while accomplishing this task