Peter Cochrane's Blog: Bean-counting just doesn't add up

Letting accountants into the engine room can wreak havoc...

Written in my car in a massive traffic jam on the A14 - a road that should have been a motorway and not a cart track. Dispatched to over my home network a few days later.

Optimising a business for cashflow can be good or bad, but in the networks business it is always bad.

If you run a business, there is nothing quite as powerful as a good accountant - with the emphasis on good. They bring a wealth of knowledge and order that is essential in any well-run company. However, if they get bored and get into the engine room they can wreak absolute havoc.

Any company run by unknowing and mechanistic minds is in jeopardy and faces a limited life. Specifically, the pursuance of policies that focus on cost reduction and a continued optimisation of the bottom line just accelerate progress towards sudden death.

  • When reporters asked Shepard [Nasa Freedom 7 astronaut Alan Shepard, 1961] what he thought about as he sat atop the Redstone rocket, waiting for lift-off, he replied, 'The fact that every part of this ship was built by the low bidder'. Gene Kranz, Failure Is Not an Option

For telecoms and data networks, that curse of numerical minds is probably more dangerous than in any other sector. It is manifest in the thinking that gave birth to traffic grooming and the destruction of net neutrality.

The world would be a happier place if we thought about cost of ownership and lifetime earnings

Rather than focusing on costs, the world would be a happier place if we thought about cost of ownership and lifetime earnings
Photo: Shutterstock

Giving one class of bits priority over another, or the traffic of one company or group over another, is the kiss of death. People, groups, companies and devices are not static in their habits or use of technologies. They are dynamic, very dynamic.

Optimising anything introduces brittleness, and the fine-tuning of networks is one of the most sensitive areas. Squeeze the last drop of capacity out of a pipe and the failures will be sudden, without warning, catastrophic and very expensive.

If we could only get people to think in terms of cost of ownership and lifetime earnings, how different and happier the world would be. Optimised systems don't cope with large peak-to-mean demand ratios, they like averages.

Don't believe me, or having difficulty following the argument? Consider power generation and supply, water storage, flood and river management, coastal defences, salt for roads and airports in winter or hospital cleanliness.

All have been subject to economic tuning that leads to power outages, hosepipe bans, floods, land erosion, traffic and airport chaos, and the spread of lethal infections. All with costs that are huge, but very difficult even to estimate.

The management argument is always couched in immediate and necessary cost-effectiveness, but the cost of a disaster is always left out of the equation.

Balancing potential risk against payback never seems to be on the agenda unless health and safety with personal culpability is an overt feature. As for telecoms and the internet, they now control everything, and I just hope we are not heading for the perfect storm.

My greatest sadness is that I see the engineers - and that is sometimes my team and I - designing for likely eventualities that are then overridden by management, only to see an anticipated tragedy strike later.

When I was educated as an engineer, economics was one of the many disciplines I was taught, and it has always been an aspect of all my designs and solutions. Sadly the converse is not true.