Keeping Critical Systems Up 99.9% of the Time - Intro

(this article is part of a series)

This series of articles will help you avoid ever having to think:

"My web server has been offline for 3 hours. My hosting provider isn't answering phone calls. I have no idea what to do."

These articles are aimed at the small online business owner. There are costs involved in providing for redundancy - but they don't have to be large. Figure on spending a few hours of your time, about $150 on monitoring software and then about $35 each month to maintain a backup web server.

Hopefully at some point in your career your website will account for a large part of your income. You might even progress to point where your application makes up nearly 100% of your personal earnings.

This makes keeping your "critical systems" online all the more critical.

These at minimum include:

Be highly skeptical of any web hosting (or other service) provider claiming they provide 99.9% uptime. For about 6-months I hosted with the now-defunct Dellhost which made such a claim. During that period of time I suffered from multiple multi-day outages. I would estimate Dellhost managed about a 96% uptime up until the time I fired them.

There's also a psychological factor. If I find myself in the situation where some part of SpamButcher's infrastructure isn't working, and I can't do anything about it - I pretty much go nuts. If I'm not selling email filtering software, I'm not happy. I can only assume fellow online business owners experience the same frustration.

The short answer to the problem of reliability is redundancy.

Expecting a single service provider to be up 99.9% of the time isn't realistic. Hurricanes and earthquakes happen. The IT guy is not showing up to work when his house is under water.

Some have even argued that a December 2006 earthquake in Taiwan took so many spamming servers offline that it caused a drop in spam.

However, if you have two different providers that are up 99% of the time - the odds of having both of them fail simultaneously is about 1 in 10,000 (that would be 99.99% uptime!). Your actual uptime is now dependant only upon how quickly you can switch over in event of a failure. 99.9% uptime is a realistic goal.

At least as importantly, if your main ISP is shaken to rubble by an earthquake - you can be back online within minutes as opposed to days.

The internet's architecture is actually designed to support redundancy. Unfortunately many business owners don't take advantage of this due to the additional costs involved.

Having backup systems in place obviously isn't free, but it's not that expensive either. If you're making over $1000 a month via your website, it probably makes sense to deploy at least some level of redundancy.

In the next few articles, I will provide specifics on ways to implement monitoring, and redundancy for your web, email and ecommerce systems.

Next article: Monitoring

Back