As the number of these trading systems multiply, the possibility of a failure multiplies.
In this case, the SIP (Security Information Processor), the network that carries the quotes and trades for Nasdaq, failed. It wasn't technically Nasdaq that failed, it was the SIP. We don't know why.
When the SIP fails like this, it shows a serious mistake in design implementation.
(Read more: Nasdaq shuts down, but stock market shrugs it off)
Somewhat surprising is the lack of a functioning backup system. Presumably, there is one (the NYSE has redundancies build into the system). For whatever reason, there either was not an adequate backup or it was not utilized. There has been an argument for some time that there is not a great need for a backup system for exchanges because all the other market centers provide natural backups.
In theory, that's true, but it's not working out that way. Trading in Nasdaq-listed stocks at the NYSE was halted because the Nasdaq declared a Regulatory Halt, which also shut down trading in Nasdaq stocks on other exchanges, including the NYSE. So much for natural backups.
Why the Regulatory Halt? The SIP distributes pricing of Nasdaq to everyone, including retail traders. It's like a consolidated tape. Without the SIP, everyone would essentially be flying blind on pricing. It was one of the worst possible components to fail.
(Read more: Cramer: We need a disaster plan, now!)
What can be done to improve trading technology? The SEC has floated a new proposal, called Reg SCI, which would establish uniform standards for technology across all the exchanges. It has floundered. It should not.
Jon Najarian made a good point on our air earlier today: we are not going back to a horse and buggy era. Automation is here to stay.
But if there is a serious problem, we should have a transportation board investigation where the industry comes together and figures out a fix. The fix would require upgrades to software at the very least.
Rich Repetto at Sandler O'Neill and others have pointed out that nothing really happened after the Flash Crash of May 2011 nor Facebook in May 2012. Hopefully this will drive larger, more concrete and visible changes.
That's priority one. Separately, there should be more serious discussions of how to simplify the trading system. Do we really need 13 different exchanges? Dozens of different dark pools? Don't hold your breath expecting a big change on this front. There are too many entrenched interests.
One final point should be made: the stock trading business is terrible. There have been layoffs in brokerage firms, and layoffs at the Nasdaq and NYSE. Staffing is lighter than it was five years ago and getting lighter. That might have to change, at least with the technology staff.
—By CNBC's Bob Pisani