Detection and blocking of security threats against organisations often is done through IP address-based methods and reputation services, but Juniper this week launched an effort to encourage security managers to abandon IP-based detection in favour of the "device fingerprinting" its security gear now supports to pinpoint devices used in online attacks. The idea is getting mixed reviews so far.
Juniper's device fingerprinting pinpoints attacks from specific devices and identifies them in a way that can be disseminated through its Junos Spotlight Secure global attacker database and shared among Juniper customers where this threat intelligence can be put to use in Juniper security products that guard web applications and other gateways.
Juniper customers Forbes and Revlon backed the approach in public statements made this week. "Current protections need to evolve beyond IP-based blocking to definitive attack prevention and we see Juniper's new products as a step in the right direction," said David Giambruno, senior vice president and CIO at Revlon.
The idea of pinpointing devices known to be used in attacks and automatically detecting and blocking them is so compelling, that Art Coviello, executive chair of RSA, the security division of EMC, alluded to the Juniper announcement during his keynote yesterday at the RSA Conference, saying RSA would be contacting Juniper to find out about possibly including this type of device fingerprinting in its own threat-intelligence feeds.
Device fingerprinting — it's not an entirely new technology by any means — appears to have appeal to security professionals though they have qualms about abandoning IP-based threat detection. And they wonder if Juniper's device fingerprinting technology might raise the same old issues about vendor lock-in.
When a panel of four chief information security officers (CISO) at the RSA Conference here this week was asked their reaction to the idea of abandoning IP-based detection in favour of what Juniper is proposing, their reaction was mixed.
Carter Lee, CSO at e-commerce company Overstock.com, said he was interested in the idea of device fingerprinting as an additional form of threat intelligence, but he was hesitant on the idea backed by Juniper that enterprises abandon IP-based detection altogether. He also expressed concern about whether device fingerprinting might be subject to vendor lock-in, as some technologies are. And he wondered about how resistant to malware attack such a device fingerprinting technology might be. "Would some malware figure out a way to defeat that?" Lee said.
Asked for its reaction to the Juniper announcement, Cisco also weighed in.
Cisco vice president of security Dave Frampton remarked the only way to make Juniper's device fingerprinting practical and effective would be to take feeds from multiple sources in order to have it scale on a global basis. Frampton also said Cisco disagreed with the notion that IP-based detection is somehow obsolete or ineffective, as Juniper appears to claim.
And he said Cisco does have its own kind of device fingerprinting but it's used to determine specifics about "the user device and the posture of that device, such as the application running on it, the server, the geo-location and IP address," and it's seen as part of monitoring devices on the move.
"We're not labelling something an attack device and publishing it out," said Frampton. He notes there could be possible drawbacks to labeling a device that way for the purposes of threat intelligence-sharing.