by Paul Rudo on 09/06/10 at 1:37 am
In 2003, the entire north-eastern part of the United Sates, and much of Ontario, Canada had been hit with a massive blackout that knocked out some of the busiest datacenters in North America. Companies were reporting outages lasting for days, and causing millions in damages.
If your company has mission-critical applications that needed to remain live 24 hours per day, would your company survive an event like this? These are important things to think bout as we sit in a time of worldwide financial and energy crisis.
The only way to truly ensure the resilience of your IT infrastructure is to have it geographically distributed across a wide area. This way, even a nuclear blast can’t take your worldwide business offline.
For larger companies with complex IT infrastructures, distributed computing may be a good option. This is one of the factors that have triggered the popularity of virtualization.
Of course, the idea of operating multiple datacenters may not be financially feasible for smaller organizations. Thankfully, the cloud has a number of cost-effective options.
On the upper end of the scale, there are companies that offer hosted “high-availability” services. These are similar to online backup service. They continuously replicate your server to an off-site facility while also monitoring your server availability. The moment your server goes down, operations will be instantly switched over to the remote server until you decide to switch back. Instead of paying for a remote datacenter, you simply rent the space from the provider.
This makes it a very cost-effective option.
The next level down in price would be so-called “rapid recovery” systems like IBM fastback. These systems are similar to online backup, but they allow you to download your data through random-access instead of downloading your recovery data in batches. This allows you to begin using your servers again within minutes…. even while the download is still in process. It’s not 100% high-availability, but it is cost-effective way to significantly reduce downtime.