The Blind Spot of the Security Industry: SSH User Keys

More than 95 percent of the world's enterprises rely on SSH user keys to provide administrators and developers an effective means of gaining encrypted access to critical infrastructure: operating systems, applications, payment processing systems, databases, human resource and financial systems, routers, switches, firewalls and other network devices. It is a lifeline of traffic flow within our data centers, our cloud environments and how our third-party vendors and supply chain access our environments. It has done its job quietly and efficiently over the last two decades. Unfortunately, the access that SSH has been providing, in particular the access SSH user keys provide, has gone largely unmanaged - to an epic degree.

There's an unfortunate parallel here to a recent Oscar-winning film based on the events that took down Wall Street in 2008: The Big Short. Iconoclastic investor Michael Burry discovered that the subprime home loans market was headed for default. He bet more than $1 billion against the housing industry - an unprecedented move. Sadly, for everyone but him and his investors, Burry was right.

Just as Wall Street had the blind spot when it came to subprime loans, the security industry has a blind spot regarding SSH protocol and SSH user keys. In fact, there are three parallels between the movie's plot and the current state of SSH management:

• Lack of understanding about the problem and its severity
• How challenging oversight of the problem is
• How significant the impact or consequences are of not addressing the problem

A Real and Far-Reaching Problem
The SSH protocol is not well understood by the majority of people, much like the subprime loan market. In fact, very few understand in sufficient detail the underlying critical access it is providing to our most important infrastructure. In this case, however, ignorance is not bliss.

For instance, in a typical financial enterprise with 20,000 Unix/Linux servers, one can expect to find up to 4 million SSH user keys providing interactive and machine-to-machine-based access. In many cases, 10 to 20 percent of these keys provide root-level access and cannot be associated to an owner within the enterprise. Root-level access is the highest level of privilege at an operating system level. It is not just a compliance and risk issue. It is an issue of resilience that has the opportunity to impact the potential downtime of critical services within operations.

Oversight and Governance Are Difficult or Non-Existent
Chiefly because SSH has long been seen as an encryption protocol rather than a means of access, this problem has gone unnoticed for a long time. As a result, it has not been considered as a part of the access governance processes and frameworks. In fact, up until October 2015, there were no NIST guidelines related to the best practices associated with SSH user key-based access.

HIPAA, SOX, PCI and other regulatory guidelines do mention access controls, such as least privilege and segregation of duties, but none of them specifically address SSH user keys as a form of access that needs to be controlled. Is this due to a lack of understanding of SSH or an unwillingness to open Pandora's Box?

In terms of the lack oversight and governance of SSH user key-based access, there are three technical aspects that must be considered. First, SSH user keys are the only form of access a user can provision themselves without oversight or control. Unlike passwords, an administrator, developer or third-party vendor can provision themselves with access to your critical infrastructure or automate a process or file transfer. This process of provisioning, de-provisioning and recertification of SSH user key-based access is infrequently if never addressed in the IAM frameworks of enterprises.

The second consideration is that SSH user keys don't have expiration dates, in contrast to certificates. This means that SSH user key-based access continues to exist, even after an application is decommissioned or a user leaves the company.

And the third consideration: an SSH user key does not need to be associated to a user account. This means a key will not necessarily establish the identity of the user associated with it. When an SSH user key is generated, the identity that associated is not connected.

Taken together, here's what these considerations mean. A user can provision SSH user key-based access themselves without oversight to my most critical infrastructure. This key-based access does not expire. It's not clear which identity it's associated with, and there are millions of these across an enterprise, which no one has an inventory of, and they are providing access to the most critical systems.

The Ramifications Are Significant
There is an additional cause for concern in addition to the lack of understanding of the technical aspects of the SSH protocol and the lack of regulatory oversight and governance of SSH user key-based access. SSH is the malicious actor's (either a rogue insider or hacker) tool of choice, and it is the primary means by which they move laterally within an enterprise to gain access to new assets and further elevate privilege, as well as how they exfiltrate data and assets from an organization. In fact, LightCyber's Cyberweapons 2016 Report indicates that in over 50 percent of the cases, the SSH protocol is the tool of choice to gain this lateral movement across the assets of an enterprise.

It's also important to factor in the reality that bad actors are aware that SSH user keys are not being managed. As a result, they go after private keys to gain access to assets. From there, they will often create new key pairs, which will generate them access to the outside directly or move those assets automatically to servers in the cloud. This is all achieved using something called "port forwarding" via "SSH tunneling," which would allow the malicious actor to extract this data via the encryption itself, rendering firewalls ineffective.

This kind of activity takes place right within our defensive structures. Why? Because we don't know who owns the keys to attain that encrypted access and because we are not able to look inside the encryption of those sessions. The potential impact is clear. We lose our data - or worse, our customer's data. We lose our intellectual property. We lose our reputation, we lose our brand and, in turn, we lose revenue and shareholder value.

The financial implications are clear here, but there's another implication that's at least as serious: operational resilience. The potential downtime in critical systems, should SSH user key-based access to those systems be compromised, is a concern we need to seriously consider. Our disaster recovery systems, which are failovers to our production systems, often share identical SSH user keys that ensure the processes of those systems fail over in an identical manner. This means that any compromised keys in production environments can also equally affect the failover to disaster recovery systems.

Clear Parallels Pointing to Danger
Let's circle back to the similarities between what's going on with SSH user key management and the events leading up to the housing crisis. The parallels are uncanny. There's a lack of market understanding around the power of the SSH protocol and the criticality of the access it provides to our infrastructures. There's a gap in regulatory and governing oversight of how this SSH user key-based access should be monitored, provisioned, de-provisioned and recertified. There's an encrypted protocol, which malicious actors use as their tool of choice to extend their access and reach within our enterprises, create backdoors and exfiltrate data. As a result, there's a potential financial, operational and brand impact that can conservatively be described as significant to our businesses.

Poor SSH user key management may not send all organizations using SSH into a rapid downward spiral that obliterates most of them, but it can still result in devastating damage to any organization that's ignored this issue.