Attivo Brings Cyber-Security Deception to Containers and Serverless
Attivo’s ThreatDefend deception platform can now enable organizations to create decoy containers and serverless functions, in an attempt to trap attackers.
As organizations begin to embrace container and serverless technologies, there is a corresponding need to secure those deployment models. On Sept. 24, Attivo Networks announced its entry into the container and serverless security market with an update of its ThreatDefend cyber-security deception platform.
The basic idea with cyber-security deception is to provide what appears to be authentic resources, in an effort to trick and trap attackers into revealing themselves. With the new ThreatDefend update, Attivo is adding in deceptions that run natively in containers and on the AWS Lambda serverless cloud service.
“What we’re seeing as a company is that companies are starting to very actively look at deception technology for in-network threat detection and what we found is that they want to have the same ubiquitous detection across all attack surfaces,” Carolyn Crandall, Chief Deception Officer at Attivo Networks told eWEEK. “We had some support before for basically some general decoys and deception out in the cloud and we’ve now evolved it to be able to support serverless and container architectures.”
Containers, including Docker containers, provide a mechanism to isolate and virtualize running applications. Serverless on the other hand, which is also sometimes referred to as Functions as a Service, enables organizations to run functions, without the need to have a running server instance.
Crandall explained that the deceptions that Attivo has built for container and serverless deployments in the cloud include a mix of decoys that appear to an attacker as though they are authentic production containers, or serverless functions. Additionally, Attivo’s deceptions for containers and serverless also include the use of embedding decoy credentials, in an attempt to lure attackers.
“In order to attract the attackers into engaging, it (the decoy) has to be authentic and mirror match the production,” Crandall said. “So we’ve done extensive development to make sure that that is true in these new cloud environments.”
With the deception decoys and false credentials in place, once an attack occurs against the decoys, Crandall said that the Attivo platform provides attack analysis and forensics, to help organization understand how attackers are operating and what should be done to stop them as part of an incident response process.
How It Works
Part of the ThreatDefend platform that Attivo is extending for the new cloud, container and serverless security capabilities is known as BOTsink, which provides network-based threat deception for post-compromise detection. BOTsink works with public cloud providers and can also be deployed in a Kubernetes container cluster deployment.
“You can deploy ThreatDirect Containers in a Kubernetes environment using any orchestration environment,” Marc Feghali, Co-Founder and VP of Product Management, told eWEEK. “The ThreatDirect conainer will tunnel back to the BOTsink platform and project all the deception functionality on that local subnet.”
Feghali explained that in an AWS cloud, an organization first deploys BOTsink into an availability zone. On top of that is the Attivo ThreatDirect component of the ThreatDefend platform, which Feghali said is basically a virtual machine, that consumes local IP addresses and presents the decoy services that an attacker might engage with.
He added that an organizations can also choose to deploy a set of decoy credentials, including account access tokens and SSH keys inside of a Lambda serverless function, and as an Amazon S3 storage bucket. Feghali said that as such, if an attacker somehow gets into a cloud deployment and is looking for account credentials in Lamda functions, they will find the decoy credentials, which don’t impact the production environment, but can tip off the organization that an attacker is present inside their cloud deployment.
“One of our assumptions at Attivo is that the attacker is already inside the environment,” Feghali said.
Feghali explained that when an attacker is already inside of a cloud environment, it’s possible to view all the different resources that are present.
Once an attacker goes after one of the Attivo decoys, Feghali said that at a minimum, organizations are alerted to the offending account, which system has been infected and what credential was used. He added that through engagement with the attacker, the ThreatDefend platform is able to provide users with the Tactics, Techniques and Procedures (TTP) of the attack.
“We integrate with other systems like NAC (Network Access Control) and endpoint security solutions to block data exfiltration and quarantine infected endpoints,” he said. “We have an extensive integration ecosystem with over 25 vendors and we’re always expanding.”
Crandall said that Attivo differentiates from others by having broad attack surface coverage for both network and endpoint deception. She also sees Attivo’s approach as being believable by attackers, such that the decoys appears to be identical to authentic services that are running in production environments. Part of the decoy authenticity is enabled through machine learning capabilities that are part of Attivo’s technology.
“This is not machine learning to detect the attacker but machine learning to prepare, deploy and operate the credentials and decoys,” she said.