After shutting the payment section of the $365 million toll road's Web site yesterday, the NZ Transport Agency admitted it had done so because of flaws pointed out to the Herald by computer experts. It said the toll payment system was set up "without all the necessary security features."
The agency has until Jan. 25 to plug the security hole, but more than 900 motorists have sent credit card or bank details over what it now admits was an insecure Internet link to set up toll payment accounts.
This closure follows a marketing blitz the agency undertook to attract motorists to use the site. Now, the transportation agency isn't sure whether or not it will be able to get the site open in time.
I can be sure of a few things out of this: It's going to cost the NZ Transport Agency more money to fix these problems now than if it had started thinking about security earlier in development. It's generally more cost effective to build secure systems from the jump than try to bolt security onto a system that's already built.
Especially payment systems. Twitter is probably going to also learn this painful lesson.
This has been going on since LANs were first installed in corporate networks. It got worse when those LANs were bridged onto wide area networks. And the application security problems got exponentially worse with the advent of the Internet. You'd think more organizations would learn by now.
But they won't. Which is why I can be sure of something else from both of these seemingly unrelated examples: These types of gaffes won't stop. Organizations will continue to push unsecured systems out the door: and they'll likely have to pay a hefty price to secure those systems later -- and so will we in continued breaches.