There wasn’t a lot of coverage of yesterday’s Globex outage in the US. Globex is one of the trading systems used to exchange futures contracts, on the Chicago Mercantile Exchange. Details are sketchy, but it went down some time after 10am and didn’t resurface. Less than a 1/3 of normal daily volume had gone through. The story is that it was caused by problems “with Internet switching technology”.
Globex isn’t vital, but there will be a lot of people caught on the wrong end of trades as a result. Some of them will have hedged, but not all. Interestingly, the Dow was headed straight down until the outage when it promptly turned around and headed back to positive territory at the close. Ditto the Nasaq and the S&P. The opening today could be interesting, depending on the morning reports (Friday is Jobs day).
This shows that technology can still, even in the most watched environment, cause serious problems. People will lose money on this outage. Some will lose big money.
So, the argument will present itself sometime in Government about whether incidents like this give more evidence that centralisation, rationalisation and consolidation are bad things and we should, instead, encourage all departments to do their own thing – that way, if there’s one failure it won’t affect many other services. This is the “don’t put all your eggs in one basket” argument, versus my version which is “put all your eggs in one basket and then watch the basket. Very closely”.
The technology to deliver the kind of services we are trying to put together – ones that join up multiple back ends across multiple departments – are inherently very complicated. The more bits of string you have, the worse the knot to untangle. But, to duplicate the string in several places and then try to thread it through various holes is not workable either. I’m not an out and out centralist – just pro putting things in the middle that would otherwise proliferate and cause pain later.
It will be fun to watch how the debate on central versus distributed progresses.