technical document ref# IT004_san_dr.html
issue date: 12 Sep 2002
by techtolink technical division

Why SAN is best for a disaster recovery plan
A resilient SAN infrastructure can help prevent outages

    Disasters are not confined to fire, flood, earthquake, or any specific type of physical outage such the events on Sept 11. In fact, such physical events represent only a small percentage of outages. It is, therefore, critically important that companies plan for and protect their data storage assets against more prevalent outages caused by hardware and software problems, as well as by the most common source of outages: human error.

    The Disaster Recovery Journal - - defines a disaster (for businesses) as 'any event that creates an inability on the part of an organisation to provide critical business functions for some predetermined period of time'.

    A company has three choices of business continuance solutions. These should be based on a combination of its acceptable level of downtime, criticality of data, and of course, the budget.

    Continuous availability: This approach is appropriate for companies which need their data to be backed up and running within minutes of an outage, such as banks, insurance companies, telecom service operators, brokerages, e-commerce firms, and others.

    Rapid recovery: This approach is best for companies which can afford a few hours to recover, where some duplication of information and the environment has been performed, such as companies which are not immediately customer-facing. It is also for backoffice departments - including finance, HR and marketing. However, some recovery processes are required before business processes and applications can be brought back up.

    Basic recovery: This approach is sufficient for companies that can afford a day or more to recover. They can use basic backup tape procedures and SAN (storage area network) based techniques such as electronic tape vaulting.

    Disaster-recovery strategies must involve contingencies for hardware and software components that can be put in place rapidly to respond to various events. The same applies to network connectivity and services. Businesses need to design networks that can automatically reroute failed nodes. This is often accomplished by plugging local connections into larger, fault-tolerant, self-healing networks.

    How can organisations reduce or eliminate the threat of data loss and minimise downtime in the face of a disaster? They should consider two issues: One, every business needs a resilient storage infrastructure that can help prevent outages-caused disaster - whether by component failures, software failures, or human error. Two, businesses must employ storage software products that make disaster recovery work. Such products must include a backup mechanism with a safety net that would support enterprise-scale backup and offsite tape management.

    For more critical systems, companies can extend the functionality of the products by adding real-time replication. They can also use software-based 'mirroring' to move data out of the data centre and into other local or remote sites.

    Getting applications up and running is also a critical component of an effective recovery. For this to happen, the software must allow the data centre to pass control to a remote, disaster-safe location within minutes. To achieve this companies need a resilient infrastructure - one which will provide the platform for remote disaster tolerance solutions which can be implemented locally or across thousands of kilometres.

    SANs are essential to making disaster recovery work. Implementing a SAN may prevent disasters - or at least help tolerate failures that could otherwise be disastrous. While the technology cannot prevent fires or floods from occurring, a resilient SAN infrastructure can help prevent outages caused by component failures, software failures, link failures or human errors. SANs can provide alternatives to scheduled downtime by allowing maintenance procedures without disrupting business functions. They can also provide a platform for remote disaster tolerance solutions over small or large distances.

    Your company's recovery needs can be met with a traditional tape backup, vaulting and restore process, or through a more advanced solution such as clustering, replication, and high availability.

    Whatever the mechanism, you should install a full range of solutions that can keep business-critical data online and available - even when there is a distant rumble.

    The writer is director (South-east Asia & India), Brocade Communications.