Chapter 10. Long-Distance Storage Networking Applications

Impressive forces of nature continually shape our planet. Over the ages, earthquakes have formed mountain ranges and floods have carved huge canyons that today attract tourists and sports enthusiasts . However, when people and buildings ”and data centers ”are in their path , these forces can be devastating.

In the early 1990s, Hurricane Andrew flooded more than three dozen major data centers. Since that time, we have become even more dependent on quick access to stored data. Most IT managers now have remote data replication systems for disaster recovery, but those systems can be costly and cumbersome to manage.

Remote data replication consists of archiving or mirroring. Archiving is a commonly used process in which data is written to portable media, such as optical disks or magnetic tapes. For disaster recovery purposes, the disks or tapes are physically transported to an offsite location and stored. Depending on the regulatory or self-imposed replication requirements, the data may be transported across town or hundreds of miles away.

Data mirroring operates over a metropolitan- or wide-area storage network and can be fairly bandwidth- intensive , as it requires either continuous replication using synchronous mirroring of disk arrays or batch processing for asynchronous mirroring. Mirroring enables two sites to share the same set of data so that either can continue to operate on that data in the case of a site outage . Due to the distance and performance limitations of the conventional Fibre Channel switch technology, mirroring had, until the advent of IP storage networking, been used mainly for critical applications and had been limited to relatively short distances.