The following article was modified on: 2010-10-01

When designing a highly available solution for Mailbox servers, you need to ensure high availability for a variety of infrastructure components, including:

  • Infrastructure services, such as Active Directory and Domain Name System (DNS)
  • Database availability group (DAG) member servers
  • Individual storage components, such as disks, storage controllers, and storage shelves
  • Individual network components, such as routers, switches, and aggregators
  • Server and storage racks
  • Power buses
  • Datacenters

Each of these component areas represents potential points of failure, which are sometimes referred to as failure domains. As a result, the availability level of your DAG ultimately depends on how you design the solution to isolate and minimize the negative effects that a failure in one of these domains can have on your DAG environment. To achieve independence between failure domains, each failure domain must have one copy of the database. In addition, because a failure would result in multiple copies being unavailable, no more than one copy is required per failure domain.

For example, consider a scenario in which you have two copies of a database. Each copy is stored on a separate set of disks but both are located within the same storage array. If the storage array fails or becomes unavailable for any reason, both copies would be unavailable. In this example, the failure domain is the storage array. Only a single copy of each mailbox database should reside on the array. Otherwise, if the array fails, multiple (perhaps all) copies of the database will be unavailable.

When planning your mailbox architecture, consider the following additional design points:

  • Will you deploy multiple database copies?
  • How many database copies will you deploy?
  • Will you have a site-resilient architecture?
  • What kind of Mailbox server resiliency model will you deploy?
  • How many Mailbox servers will you deploy?
  • What backup model will you utilize?
  • What storage architecture will you utilize?

For detailed information about how to plan for these questions, see Understanding High Availability Factors.

Contents

Unbalanced Database Copy Layouts

Designing a Balanced Database Copy Layout

Active Database Distribution in Example Scenario During Server Failures

Design Scenarios

Looking for management tasks related to high availability and site resilience? See Managing High Availability and Site Resilience.

 

To read more on this DAG Article, please visit the next link :

 

Database Copy Layout Design: Exchange 2010 SP1 Help
http://technet.microsoft.com/en-us/library/ff973944.aspx

 

Kindest regards,

 

João Ribeiro