Today’s news around the New York Stock Exchange and the Wall Street Journal going down, as well as a glitch at United Airlines that grounded all flights this morning, are leading people to suspect a terrorist cyber attack. As of now, NYSE and the White House are both denying a cyber attack, and so far, it seems that it might just be an “internal technical issue” as NYSE stated earlier today. But absent any real explanation, it’s easy for minds to wander to a Hollywood-style super villain attack that is the plot of way too many movies. But it’s true that the financial markets are a prime target for cyber-criminals who want to cause a lot of panic, whether for politically-motivated reasons or just for bragging rights. Our financial system is so dependent on computer systems to track, manage, and transfer monies, agreements, stock trades, and account information, that any outage or degradation could have wide-spread effects.
What do I believe? Well, I think it’s definitely possible to have three newsworthy events happen on the same day – as our computer systems grow more complex, it’s also easier for them to go down, and not necessarily through a malicious attack. For example, upgrading the software on a massively complex system poses significant risks – for systems as large as NYSE for example, testing a software upgrade or patch on some of their main applications could be hard to fully test. In a lab or controlled environment, upgrades may go smoothly, but for a system that may have thousands of endpoints, gigantic data repositories, and interconnections that span multiple disparate systems through various protocols, upgrades are scary.As far as mitigating or reducing risks for these complex systems, there’s no single answer. It’s a plan and a process that need to be thought through starting from the initial system design requirements, to the overall system architecture, to a solid and robust technology stack. Also, operationally, systems should be modularly constructed so it’s easy to couple/decouple components and change/upgrade/replace a small portion without affecting the other parts of the system. A fully redundant system or at least a backup system could also be deployed, but this of course could add significant cost (although, what will be cost to NYSE due to this outage?) And of course, security is critical – it needs to be a theme that woven through every single process, connection point, interface, and storage repository – end-to-end security is the goal.
It will be interesting to see what ends up being the root cause of these outages at NYSE, UAL, and the WSJ – hopefully it’s a simple technical glitch, and not a terrorist attack. But what will these companies do to address the issue in the future? What changes will they make? Those are the interesting questions.