I remember going to Telecom ’91 in Geneva when I was working at Unitel. The fall of 1991 was the golden era of telecom exhibitions and my team was being courted by various suppliers for our network technology business.
At the time, there were strong personal relationships that linked Unitel’s CEO with the head of Digital Equipment. We were being given a special tour of the multistorey DEC booth at the show when Digital’s Canadian CEO commented to me that 100% of DEC’s Canadian data network was running on Unitel’s facilities. Without skipping a beat, I said that if he had mission critical applications running, he should fire his CIO and immediately get some carrier diversity – at least 10%. Besides, it helps keep both service providers on their toes. It makes everyone more responsive.
My CEO nearly dropped the mini-cigar from his mouth.
I explained that it doesn’t matter how good we are, networks can fail. When that happens, do you really want all your eggs in one basket?
That was a little more than 20 years ago.
Think of the types of network failures that have happened through the years for all sorts of reasons: a fire that knocked out service to downtown Toronto with repercussions across the country; a satellite failure that isolated Nunavut last fall; and this week’s fire in Calgary are just a few examples.
It isn’t just physical damage that can knock out a network – many naive architects think that they can design around single points of failure, such as fires, satellite failures or fibre cuts. Sometimes it is a matter of software, such as Amazon’s network failure last year. Given the complexity of network systems, it is difficult to laboratory test every possible condition. A software upgrade can sometimes bring down an entire network, such as that experienced by AT&T in 1990.
These are some of the issues faced by everyone who tries to operate carrier-grade networks. Service outages are a fact of life; minimizing the impact and quickly recovering from the problem is the primary objective of everyone involved.
As an end user, carrier diversity provides improved reliability.
As public safety agencies move forward with making use of their spectrum, it seems to me that reliability and overall capacity can be enhanced through interoperability with commercial networks.