Last week we heard of yet another egregious security breach at an online provider, as crooks made off with the names, address, and birth dates of eBay users, along with encrypted passwords. They suggest you change your password, which is likely a good idea, and you better also change every other place you used the same password. But that's not much help since you can't change your name, address, and birth date, which are ever so handy for phishing and identity theft.
According to eBay's press release, crooks
compromised a small number of employee log-in credentials, allowing unauthorized access to eBay's corporate network, ...
and then vacuumed up 145 million users' credentials. Say what? Random eBay employees can access all that customer data from the general corporate network, and all that's keeping it safe is the hope that none of their 33,000 employees have had their passwords phished? No wonder several state Attorneys General have opened investigations.
I do not purport to be a great data security whiz (knowing, as I do, just enough to play one in this blog), but even I know that you don't hook the valuable data directly to the corporate network, you limit bulk access with hardware tokens, and routine access with APIs that provide just what they need to run the system, e.g., verify whether an entered password is correct, retrieve one user's e-mail address so they can send her a note saying she's been outbid. You use multiple levels of security so if one breaks, the damage is limited by the others. All of this stuff was old in the 1980s. What were they (not) thinking?
We need to ask the same questions about AOL's giant data breach and Yahoo's giant data breach (not to be confused with their previous giant data breach in 2012), and Target's giant data breach and likely a dozen more that the companies haven't gotten around to noticing.
Large collections of personal data are valuable. Crooks will try and steal them. This is not news. It's 2014, so why are companies using security approaches that weren't really adequate 20 years ago for databases three orders of magnitude smaller than today's?