People and machines together can be greater than the sum of their parts, and this is exemplified by SentinelOne, a Silicon Valley-based firm that secures endpoints, datacenters and the cloud.
Vigilance is the name of the game, and also happens to be the name of SentinelOne's new service, based on its existing endpoint security platform, which augments the power of computer threat detection with a team of human analysts and researchers based in Mountain View and Tel Aviv.
The platform detects the threats. The humans examine, discuss and then respond to them. Simple.
According to SentinelOne this is an optimal arrangement where detection, prioritization and responses are accelerated because there are more bodies and know-how on the job, apparently thereby reducing corporate risk. The Vigilance service provides the elasticity to call on additional eyeballs during a high threat period, or to provide more expertise in security analysis and research than might be found in a single enterprise team instance.
There's a strong reliance from the Vigilance team on the platform to provide primary threat information -- to be the eyes and ears -- and for the analysts and researchers to be the brains. "The service ingests threat information from the agent detection, and 80% to 90% of the analysts' tasks are based on that information," said Eran Ashkenazi, VP of services and field operations for SentinelOne.
Oddly enough, the more problems the SentinelOne platform needs to handle, the better. It learns how to handle unique incidents and then that knowledge is propagated to SentinelOne's entire customer base, theoretically lessening the impact a new threat can have and spreading the benefits. At the 50,000ft level, the platform's main role is to differentiate between false and true positives, and then hand off that information to the human team. But what happens if the humans then make mistakes?
"There are several safeguards, but in short if there is a doubt, issues will be escalated to a second tier of malware researchers or reverses, or we'll interact with the customer to learn more or get the actual file or payload," Ashkenazi told SecurityNow.
An interesting aspect of the service is the ability for enterprise security teams to bring online extra personnel when they're needed. In some cases, those companies may simply have underestimated how many members of a team are needed to handle threats. But also, the sheer number of tools out there can cause headaches too.
"There are simply too many solutions to manage, and an average security team needs to probably deal with dozens of different platforms and dashboards," said Ashkenazi. "[Plus] the endpoint was for many years considered to be something that just works, which we [now] know is not the case."
In the last five years, endpoint protection, and endpoint detection and response, platforms have become more resource-intensive as the range and number of attack vectors increases. Like bacteria on a petri dish, the threats multiply in number and diversity until the whole lab is crawling with infectious organisms. Yet the number of scientists and technicians remains the same.
In SentinelOne's lab, they're looking next to incorporate AI into the platform, once deep visibility capabilities have nee incorporated, proving more data into the Vigilance service, enabling proactive hunting capabilities, and making the security stance of the platform more proactive.
- AI Just Gets Bigger
- Automation Deserves Skepticism
- Magical Thinking Drives the Myth of AI Solving Security
— Simon Marshall, Technology Journalist, special to Security Now