The Secret Third Step to Threat Detection and Response: Protection

Before threat detecting/monitoring identities for suspicious or anomalous access and remediating incidents after they happen, there’s a grey area: properly configuring your setup and taking preventative steps to protect your identities (and, therefore, your critical assets). How are you protecting your accounts before an incident can occur (or slowing an incident down before it really ramps up)?

While there are many ways to authenticate (MFA, SSO, username/password), the reality is that, like with CircleCI, we’ve seen attackers circumvent these measures more and more, and “traditional one-time gating mechanisms (once you’re in, you’re in) are insufficient to cater to this new reality,” according to The 2022 Gartner Report. In their recommendation, “trust and risk must be assessed at every moment, and changes must be reflected immediately, across all sessions. For this reason, IAM is well-placed to deliver a more continuous approach to support an adaptive zero trust security capability.”

MFA was originally a “nice-to-have,” which then became mandatory, and now it’s both mandatory and insufficient (e.g., one can steal a 2FA-backed Okta cookie and have access to anything in the SSO provider that doesn’t require SAML app sign-in time step-up). Also, as we briefly discussed in a previous article, device trust is another measure that isn’t impenetrable and can sometimes be spoofed.

In addition to MFA not being robust enough against new attacks, many find it annoying and bothersome. Most people are comfortable with multiple step-ups, like a required 2FA re-prompt + a Yubikey every time you log in, for high-sensitivity assets (e.g., AWS root). However, for lower-sensitivity-but-still-valuable assets, especially those used often, it may not be worth the social capital (from the security/IT teams’ perspective) to require additional 2FA at every app log-in since the rest of the company may find it incredibly inconvenient.

The crux of protection is this: the ideal protective state is to verify that the user is who you think they are at every step of an application usage lifecycle—from device sign-in to SSO-provider sign-in to application sign-in to application usage—in order to mitigate account takeover or malicious insiders (e.g., Account Takeover: “Sam shouldn’t be accessing Salesforce right now, they’re on parental leave” or Malicious Insider: “I don’t think it’s a good idea for Jordan to export Snowflake data right after they handed in their two weeks notice”). You want to protect critical assets (sensitive data that we aim to safeguard, like financial, customer, and business data) while still maximizing productivity.

Yet, as we demonstrated, you’ll run into tons of roadblocks with getting users to adopt measures, especially if they disrupt their workflow (not just ”ugh, why do I have to keep hitting the 2FA prompt on my phone every time I open my laptop” but also MFA-fatigue attacks, where the attacker spams MFA prompts until the user lets up, or a breach where the user just lets an MFA request through because they’re annoyed by MFA). This makes these comprehensive measures easy in theory but difficult in practice due to the operational challenges of using them.

Gartner provides a variety of advice on how to protect yourself when you suspect a threat, including “use SSE to provide a containment layer for SaaS apps,” “freeze all automated provisioning,” and “use automated threat containment approaches, such as risk-based adaptive access (step-up authentication and session termination).” Yet, in a sea of potential solutions, it’s hard to know which measures are both effective and right for you.

An important context here is company stage. For example, a small, low-profile company might have fewer threats to defend against, so a basic IdP setup with mandatory MFA at sign-in time will take them far. On the other hand, a high-profile international company (e.g., Apple, Google, Meta) has a massive target on their backs, drawing highly sophisticated attackers that will find and exploit even the most minor gaps in protective measures. For these companies, a more robust security approach would be needed.

Another contextual framework is the “crown jewel assets” model, where, as opposed to company size, company data is what sets your risk factors apart. Take, for instance, a meditation app company that deals with relatively little consumer data; this lack of “valuable” data means they may not be a big target for attacks. On the contrary, a neobank whose data includes customer financial information and bank account access is likely to be a bigger target that requires a more involved security strategy due to its perceived value to attackers.

No matter your risk level, every organization needs an information security strategy. Attacks can’t be the first time you do something about security because, then, time is of the essence (a concept we elaborate on in “Cybersecurity Is More Critical Than Ever, and You (Yes, You) Can Do Something About It Now”). Every second you spend detecting and defending against threats while a breach is happening is time that the attacker will be using to gather information, attempt to install persistence mechanisms, and potentially compromise additional accounts. This makes it so that automating protection and creating “speed bumps” like MFA and peer approval are invaluable measures to limit the costs of attacks and their remediation, even if they can’t always prevent them entirely.

While there are boundless considerations when it comes to protection (e.g., configuration drift, grandfathered exceptions, supply chain risk, etc), ultimately, everyone’s goal is to prevent and minimize the impact of an attack as much as possible.

Luckily, threat protection and detection software/solutions are starting to appear across the identity space, and we’ll talk more about them as they’re released. To get notified when we post more about these solutions (and to stay up to date with Crosswire on all things identity and infosec), sign up to receive our updates below!