Anyone who relies on computers (or any other machine) has to face the fact that "things break". But what happens to your web site when something breaks?

Many providers will gladly describe in detail their redundant networks, and maybe their redundant disk arrays. Sure, it's easy enough to cover these common and obvious points of failure, but what about the processor, or the network interface, or memory, or the motherboard chipset, or...

And what do you do about software? It breaks, too.

And what about system maintenance? That's the real challenge. Some maintenance, especially things like applying operating system security patches, can't be done on a running server. Meaning scheduled downtime.

So what do we do that's different? Our hosting environment is based on the concept of redundant servers. Not just redundant components in the server, but a whole other machine, ready to go, with all your data just standing by, waiting for a problem.

Simple concept, but not necessarily simple to implement. Here's how we do it:

First, we only load each server to a maximum of 50% utilization. This accomplishes two things. Obviously, it ensures snappy response times during normal operation, since the servers are only lightly loaded. But it also gives us spare capacity on each machine to allow it to take over for a failed machine.

Second, we pair up our servers. Each server maintains an identical copy of all data needed by both servers. Same email, same web pages, same database content. All automatically and dynamically kept in syncronization.

So when (not if) a server fails, or partially fails, we simply move the affected addresses to the standby machine. Since the standby server is already booted, already running all needed software, and already has all the content of the failed server, this takes only a few seconds or less.

And here's the real plus: We don't have to wait for a failure. Since we can switch over at any time without affecting service, we can take any server down at will for maintenance.

This allows application of security or other patches in an extremely timely manner, which actually prevents outages, since most outages are caused by software failure, not hardware failure. Using this feature, we can realize our unique goal of installing applicable security patches within 12 hours of release -- which helps meet our ultimate goal of keeping your website up!

Have more questions on how we keep your website running? Just drop up a note.

Omd3.com Hosting Services
info@omd3.com