TrustDefender debuts web-page fingerprinting in bank fraud fight
Australian security firm TrustDefender is expanding into fraud detection with the release of software designed to spot banking Trojans that manipulate web sessions.
Variants of the ZeuS Trojan and other strains of malware use tricks such as installing phoney dialogue boxes when users log into online banking sites from malware-infected machines.
These so-called man-in-the browser attacks are designed to fool marks into handing over confidential data to fraudsters, such as bank card PINs or one-time login codes. Traditional antivirus software often struggles to quickly identify and block such tactics, which fraudsters employ as a means to get around two-factor authentication for bank logins.
TrustDefender Zero detects these attacks by comparing the content served from a server to the content as seen by a client to detect whether other components of a web page have been added along the way, a probable sign of malfeasance. Andreas Baumhof, CTO and co-founder of TrustDefender, told El Reg that the benefit of the approach is that it is device independent and works without the need to run blacklists or updating.
TrustDefender Zero also includes a device fingerprinting feature that is calculated on a user's PC but not stored there. The firm is looking to sell the technology to banks and e-commerce providers who would offer it to consumers as a means of minimising the risk of fraud.
The product adds to TrustDefender's existing TrustDefender Pro end-point protection software. This technology bundles a security dashboard (that warns end users if their anti-virus is not up to date, for example), whitelisting and memory forensics (to detect the presence of rootkits on client machines).
The software works by disabling any malware that might be present on a machine during the time an online banking transaction might be taking place.
Transaction security firm Trusteer tries to solve a similar problem with a slightly different approach, using client software that hooks into browsers. Trusteer has already signed up a large number of banks in Europe and the US while TrustDefender is at the early stages of its global expansion into these territories.
Ted Egan, CEO and co-founder of TrustDefender, said banks need a choice of software providers. "We're aiming to mix it up a bit", he said. TrustDefender is also targeting e-retailers, airlines and gambling sites with its technology, but its main focus (for now at least) is online banking.
In addition to unveiling TrustDefender Zero, the firm is also launching TrustDefender Central Intelligence Server. The technology, which bolts onto TrustDefender Pro, would sit in a bank's data centre and act to either warn customers or make them go through additional security checks if malware is detected on their systems.
These extra checks might involve sending a one-time login code in the form of an SMS message to a customer's pre-registered mobile number, for example. ®
It's a mostly-futile arms race...
As soon as there's something untrustworthy in the channel between the bank and its client, all bets are off. No sort of checking is infallible if the trojan software can modify the input to and/or output from the checks. And since trojan writers aren't constrained by release cycles, beta-testing and customer acceptance procedures, they can usually turn round a response to the latest threat (as they would see it) rather sooner than a legitimate developer.
Banks are desperate for a "magic bullet" to solve the trojan problem because their customers (with their nasty infected insecure PCs) are resistant to the incovenience of out-of-band verification and banks are resistant to having to issue and support additional PC hardware like smart cards. They're likely to remain desperate unless they start to have some serious conversations with each other and with the browser developers.
js-based diff as malfeance detection tool
Well, as a tool it looks fairly useful. As a workflow enhancement it's that more popups. Dashboards that let you keep an eye on all sorts of things that you really shouldn't need idem dito. They're useful more from an engineer's viewpoint than from a user's viewpoint.
Also likely useful mostly from an engineer's viewpoint would be something that made transparent just what came from where on the page. Including pictures and other content from a wildly different domain is /de rigeur/ these days, but it does little but subvert the notion that what you see on the screen is from where the adress bar says it is from, too.
How to translate all that into something truly useful for DAUs, avoiding creating yet more GWF type trouble, is something the various "it security" firms should concentrate on. But they're not doing it, at least not that I can see. It's really quite depressing how unimaginatively the IT security business innovates.