• United States

How many layers of redundancy are enough?

Jan 24, 20063 mins

* Spreading your risk to ensure uptime

If re-examining your network redundancy plans was one of your New Year’s resolutions, it may turn out to be just in time. We’re barely past mid-January, and networking entities of all sizes are quickly discovering that having “enough” independent paths may involve a larger number of paths than anticipated.

We’ll start with a relatively massive network outage that occurred for Sprint customers in the Western U.S. on Jan. 10. According to a report in CNN, Sprint suffered an outage of local, cellular and Internet services in a large region due to a fiber cable cut between Phoenix and Palm Springs. And why was the traffic not rerouted? Crews were already repairing a cut at another route that would have been used as the alternate path.

While it’s true that having two physically diverse paths usually provides great backup, the more paths the merrier. For instance, assume that you have two separate paths, that each path by itself has 99% availability and that if one path fails, there is no impact on network performance because the other path would handle all the traffic. In this case, the availability – because you only have to have one of the two paths available – goes 99.99% (four nines). But if you then add a third path with 99% availability, you’ve increased your availability to 99.9999% (six nines).

These calculations are what convinced Steve several years ago that it was a good idea in the SOHO environment to have both cable modem and ADSL service. After all, they even have diverse routing to the premise. But he also got caught with sub-optimal service a couple of weeks ago when there were independent problems on the two services. The cable service was having a high error rate and lots of modem resets, while the ADSL service was not able to reach his main e-mail server.

By the way, the inability to reach the e-mail server was not an ADSL problem per se. It was, instead, a horse of a different color in which the reliability is based on multiple independent systems all having to be in service.

In a future set of newsletters, we’ll discuss the “magic” behind availability calculations. In the meantime, Steve is thinking seriously about calling the phone number for the new wireless provider in the neighborhood, potentially providing a third independent path, although he is greatly embarrassed to go so far as to contemplate responding to one of those hanging ads left on the mailbox.

Jim has a broad background in the IT industry. This includes serving as a software engineer, an engineering manager for high-speed data services for a major network service provider, a product manager for network hardware, a network manager at two Fortune 500 companies, and the principal of a consulting organization. In addition, Jim has created software tools for designing customer networks for a major network service provider and directed and performed market research at a major industry analyst firm. Jim’s current interests include both cloud networking and application and service delivery. Jim has a Ph.D. in Mathematics from Boston University.

More from this author