The more connected the elements of a system are, the greater the robustness of the system, right? If some of the links break, there will be alternative paths; so the system can keep going while the broken links are repaired. That is a favorite mantra of Internet evangelists; and the principle is so simple that it is an easy matter to get everyone else saying the same thing (sort of like hearing "AFLAC" whenever a duck quacks). James Burke had another perspective though; and, while it was not the primary thesis of his Connections television series, it leaked out very early in his exposition. Burke pointed out that rich connectivity could induce fragility and vulnerability as readily as it could induce robustness. In other words assessing the consequences of rich connectivity is a bit like what I have called the "TBD of Crowds" problem: Whether or not a crowd is wiser or madder than any of its individual members depends on a bundle of subtle factors that James S. Coleman called collectively the "micro-to-macro problem." Thus, while the Internet may be robust enough to allow me to matriculate in a seminar on Marxist phenomenology (I thought I was making that up, but I just got 203 Google hits for it!) conducted at just about any university in the world, it is also vulnerable to a neo-Trotskyite infecting all participants of that seminar with malware, which will then propagate through every electronic mail message they each send. Needless to say, when we descend from the ivory tower of a philosophy seminar to the everyday world of work, things can get much more serious.
So it is that Al Jazeera English has compiled the following story from their wire sources:
Internet disruption effecting users in India and the Middle East looks set to trouble India's lucrative outsourcing industry and could in turn impact international businesses that rely on it.
India on Thursday struggled to overcome internet slowdowns and outages, which came after two undersea cables off the coast of Egypt were damaged.
The incident has halved India's bandwidth.
Outsourcing firms, such as Infosys and Wipro, and US companies with significant back-office and research and development operations in India, such as IBM and Intel, said they were still trying to asses how their operations had been impacted, if at all.
Rajesh Chharia, the president of the Internet Service Providers' Association of India, said companies that serve the east coast of the US and Britain had been badly hit.
"The companies that serve the [US] east coast and the UK are worst affected. The delay is very bad in some cases," Chharia was quoted by the Associated Press as saying.
"They have to arrange backup plans or they have to accept the poor quality for the time being until the fiber is restored."
That last sentence is the real kicker. It reminds us of just how naive most of us are about the need for precautionary measures, which I have previously associated with our fear of formulating worst-case scenarios. Furthermore, as the final paragraph of the Al Jazeera report reminds us, it is not as if we were totally unaware that problems like this could arise:
Such large-scale disruptions are rare but have occurred before. East Asia suffered nearly two months of outages and slow service after an earthquake damaged undersea cables near Taiwan in December 2006.
From a technical point of view, the problem may be that we have overloaded our trust in connectivity by overlooking the need for distributivity. It is all very well and good that India has become a "global hub" for providing service; but, whether or not you think the idea of having your customer service call routed half-way around the world is a good thing, why should there be a "global hub" in the first place? Why can the service burden not be shared by several sites distributed around the world, perhaps equally spaced according to time zones? (I know of at least one major business that has set up their software development centers in this way. You can have three sites, each with a unique eight-hour shift on a "global clock;" and, at the end of each shift, the work gets passed over to the next time zone.) In this case, if an entire site lost its connectivity, the other sites could extend their shifts to cover for it.
Of course we know the answer to that first question: A bunch of astute Indian entrepreneurs saw how they could turn their "home turf" into a "service center for the world," went for all the marbles, and succeeded in getting them. Probably they could have been just as good at designing, implementing, and running globally distributed service centers; but the local model may have been more profitable (possibly with government-based incentives factored into the balance sheets). So now, when a fiber cable off the coast of Egypt "catches cold," the whole world of Internet-based service sneezes (worse yet, probably all at the same time, sort of like all the toilets in the United States flushing during Super Bowl half-time)!
It has been a while since I have ranted about the fact that, for all our hollow words about being in a "service economy," we (and that, unfortunately, is a global "we") are far more ignorant about the nature of service than we have ever been about manufacturing, even when most of the world was drunk on the worst excesses of Frederick Taylor's "principles of scientific management" (which, I have also argued, seems to be the lens through which we try to view service). This is not to imply that understanding the basis of your economy makes you good (and profitable) at doing it. However, that understanding provides a good start when it comes to being aware of the problems you are likely to encounter; but, if we are going to be phobic about worst-case scenarios, will that really matter? So the next time you want to rant about the lousy service you are getting, just remember my own methodology and think about how we all got into this mess!