How to plan for SharePoint Disaster Recovery

To plan for a disaster recovery there needs to be clear business requirements for two critical factors:

• Recovery Time Objective (RTO). This is the time needed for the environment to be available. For example, two hours for the system to be back online.
• Recovery Point Objective (RPO). This is the amount of time where it is affordable to lose data.  For example, if the RPO is two hours then there cannot be any loss of data entered prior to two hours before the disaster event (but any data within the last two hours after the event may be lost). 

Once we have clear RTO and RPO requirements we can start building the strategic plan to achieve the Disaster Recovery goal.  The Disaster Recovery plan needs to include:

• Business continuity procedures (very well documented).
• Regular exercises to simulate Disaster Recovery and the results from those exercises to improve the plan.
• Up to date contacts lists for the teams involved. 

Methods of SharePoint Disaster Recovery

There are three methods for recovery and based on your requirements, you need to decide which type will be best for your business: 

1. This is where a secondary data center can provide support in hours or days.  After a disaster happens, the process of building servers and restoring shipped data from different Medias will start, which will take time. This is the lowest cost method.
2. This is where a secondary data center can provide support within minutes or hours. With this method, virtual machines are utilized. A snapshot is taken from the primary data center servers and database backups are taken. The secondary data center is frequently updated with those backups and snapshots. This method saves the time it takes to rebuild the whole farm. It costs more than Cold and less than Hot.
3. This is where a secondary data center is available and can provide support within seconds or minutes. A redirection will happen on DNS from the primary data center to the secondary data center. This is the most expensive method as the data is continually synchronized real-time between the two data centers.