Last week I spent two days in St Andrews (Scotland) with the folks from the Large Scale Complex System (LSCITS) initiative and must say that I found it extremely insightful. LSCITS is a UK cross-university initiative with a deep interest in socio-technical issues and broad base of both academic and industrial input. As such they deserve genuine credit for their independent and objective efforts into such an important area of research. In particular I must thank Erick Hollnagel for his epiphytical presentation. In this he pointed out two things of particular relevance: More and more contemporary technical architectures overly concentrate on the technical components involved and underplay the human element present. This can cause real problems when systems fail, as handing back the operation of underlying processes to human control is often designed as a single step activity. This may be fine for simple, small scale systems, but when planet-scale systems are the order of the day such immediacy is extremely dangerous.
Most often technical architects labour to identify and accentuate the benefit returned by a solution and again, historically, this has proved fine for smaller systems. However when scale and complexity ramp up, working with such an eye becomes much harder as design goals turn into more difficult beasts to control.The opposite tends to be true with failure mode analysis, however, as it is becomes easier to define what must not happen when a highly complex system fails as you get closer to it. Take the recent financial crisis for instance. This provides a perfect example of a large scale, complex socio-technical system, and its recent failure was indeed spectacular. When times were good it was the job of many highly paid individuals to squeeze every last drop of benefit from this already overburdened system. For them every day was a matter of increasingly hard work. Even so, ask any one of those individuals burning the midnight oil at that time what should happen if the system they were nurturing should fail and the answers would come thick and fast – not least of which would be that banks should invoke suitable self protection mechanisms so as not to fail!
Both of these points indicate a way forward for large scale complex socio-technical systems design and perhaps further provide the seeds for a new socio-technical design method. By doing so they essentially turn conventional thinking on its head and demand that IT Architects predominantly design from a pessimistic standpoint, thereby understanding full well that probability is working against them and failures are more likely than not. They further stress the role of wetware in any given solution and emphasis the essential characteristics of human intelligence in the ultimate success of any socio-technical system. For some IT architects this will be heresy, but for others with a more rounded background, it will be nothing more than obvious. Other, less technical, branches of systems engineering have taken such a standpoint as read for some time. Manufacturing science, for example, has held Failure Mode Effect Analysis (FMEA) close ever since its humble beginnings. Now it is clearly time for such good practice to filter across into mainstream IT architecture.
I travelled to St Andrews with my good friends and fellow Executive Architects from IBM, Richard Hopkins and Kevin Jenkins. Some years ago they both invented the Brownfield method of systems development and I am pleased to report that the car ride back was filled with feverish excitement as they discussed extending Brownfield squarely into the domain of large scale socio-technical systems. When (rather than if) this happens it will be a significant, and greatly needed, step for the IT industry and I’ve already held my hand up to help wherever I can. If you would like to follow Richard and Kevin’s progress, please see here.
Recent Comments