TrustSphere says its TrustVault product helps crucial emails get through–even in the midst of a denial of service attack–by correctly identifying trusted senders.
As annoying as spam is, an overactive spam filter is almost worse when it prevents important messages from getting through.
A company called TrustSphere says the TrustVault product it introduced this week can act as a counterweight to the spam filter, using a type of “social graph” to identify trusted senders and allow their messages to get through–even in the midst of a crisis such as a distributed denial of service attack on an executive’s email account.
“Inside the the organization, we’re effectively mapping who’s speaking to whom and turning that into an enterprise social graph,” Manish Goel, CEO of TrustSphere, said in an interview. “We’re tracking who’s speaking with whom and how often–what’s the cadence of communication.” In that way, TrustVault can identify the trustworthy senders and allow their messages to go through, even if they would otherwise be blocked by a spam filter.
So far, this social graph is based entirely on the exchange of email, although TrustSphere is working on ways of integrating social media and voice over Internet protocol communications for a more complete picture, Goel said. But TrustSphere is applying elements of social networking theory such as Dunbar’s number, anthropologist Robin Dunbar’s concept that humans can only track a limited number of relationships, often theorized as about 150, and rely on “circles of trust” for more extended relationships. In this way, TrustSphere models trustworthy connections at the organizational level, as well as at the individual level. TrustVault is also linked to a related service, TrustCloud, which tracks the reputation of email accounts across the Internet.
TrustSphere doesn’t filter the content of the messages at all, looking only at the pattern of communication and touching only the email header fields, Goel said. The service does detect email authentication methods, such as the use of Sender Policy Framework tagging, but it’s counted as an indicator of trustworthiness rather than a final verdict, he said.
Messages cleared by TrustVault can still go through anti-virus and spyware scans, and even previously trusted senders can be screened out if they start exhibiting suspicious behavior, Goel said. But sometimes letting the right messages through can be as important as keeping the wrong ones out. For example, corporations targeted by activists or hactivists sometimes have the email accounts of top executives rendered useless when they are flooded by messages sent by angry consumers or generated by bots. With TrustVault, the messages from known senders could be delivered to the executive being targeted, while all the rest would be routed for review by an administrative assistant.
One of the company’s oldest customers, the doctors.net.uk social network for physicians in the U.K., has been using a version of the same technology to allow email that uses words like “Viagra” or “penis” to get past spam filters when those words are used in a legitimate medical context, rather than for spam or pornographic promotions, Goel said.
“This also allows you to turn up the threshold on the aggressiveness of your spam filters without missing messages,” Goel said. “I liken this to why cars have brakes–to allow you to go faster. Spam filtering is very much focused on identifying the bad guys. We’re using the good and the bad to improve the overall security infrastructure.”
Founded in Singapore, TrustSphere is just now bringing its product to the U.S. market.