Today made for an interesting read in the papers, as many articles related to the Chinese stock market decline were posted and commented on by people around the country.  At one point, an article on The Globe and Mail said:

In addition to the exchange’s woes, data providers that carry buy and sell orders to the NYSE failed to keep up with selling programs in place at brokerage houses and institutional investors such as hedge funds. Dow Jones & Co. Inc., which runs the benchmark American index, saw its computers stretched past their limits, with the speed of the decline outpacing the index’s ability to track falling stocks.

From a developer’s standpoint, this is pretty cool.

The stock exchanges in operation around the world have some of the most sophisticated computer systems on the planet.  Not only are the servers incredibly powerful, but the databases and data pipes in use are so engrossing that I can have wet dreams just thinking about them.

Elaborate data structures.  PIT (Point-in-Time) Databases.  Fat fibre channels transporting gigabytes worth of financial data every second.  I couldn’t imagine having the luxury of building one of these systems from the ground up.  It wouldn’t matter where my office was … be it a tiny cubicle shared with someone who never bathed, or an elaborate corner office on the 40th floor of an elaborate sky scraper.  Constructing a system capable of handling the demands of a stock market (espescially the NYSE) would be a worthy challenge.

I remember reading several articles a while ago about this electronic system that was put into place, and it was made to handle more work per second than my employer handles each quarter.  The people that developed the systems in place were incredibly talented and it shows in the background.

So to hear that it slowed down to the point that some organizations had to throw more servers into the fray … that’s just amazing.

In the coming years I’m sure that these systems will be scaled up a bit to handle such market corrections.  Though the question would always remain; just how big of a market change could happen before the systems couldn’t keep up again?  Fortunately, this is not a question I will ever need to answer.  It would be an exciting project to work on, though.