With more focus being placed on rapid recovery times for disaster recovery (DR) operations, much of the design, strategy, and practice work done for DR in the past has shifted more toward the high availability (HA) concept. For many businesses, an “always on, 24/7/365″ concept is key, so a recovery time of 48 hours is simply too long, and a data loss of an entire week would be catastrophic and considered a definite disaster in its own right. So, availability is now king–how do we achieve it? See my article on Virtualization, Replication, Storage and High Availability for introductory concepts on replication and how storage requirements increase, and on the general ideas behind clusters and replication.

Many of you here are from a Microsoft Exchange and therefore a Windows Server environment. While much has changed in the capabilities for Windows server clustering, especially in the Exchange area, many of the core concepts are the same regardless of what the latest features and options are. For example, block-level replication across drives on a SAN solution such as EMC’s SRDF/CE option is specifically designed to assist in replication of Windows databases such as SQL and Exchange, but the block-level replication works in essentially the same manner as DRBD does on Linux.

Continue reading Considerations for High Availability Designs Used for Disaster Recovery