Identity In A Zero Trust World
Our solutions must continually validate the identity of the person accessing our applications and data.
We have a concept at SailPoint: identity at the center of security. While that may appear self-gratifying, when you dig a little deeper, you’ll discover why it’s accurate. Zero Trust has been the growing buzzword and theme for the last couple of years in security. At its core, it means we build our solutions to habitually question who someone is and what they want to do. We can no longer blindly trust a user that comes from a known location. We need to keep asking questions.
At the center of those questions is a critical concept: identity. Our solutions must continually validate the identity of the person accessing our applications and data. That includes authentication, authorization, and one that might not entirely be as obvious is administration. The identity becomes the Golden Key to ensuring that we can validate users. We need trusted sources that tell us who someone is and what access they should have. From there we can form the foundation of our policies that determine how things are accessed.
While it all sounds great in theory, what does that mean in reality? What steps do you take to get there? What does that mean for my firewall?
Let’s answer all of those questions.
First, let’s breakdown what it means to have a Zero Trust model. There is no such thing as an internal network. Whether someone is accessing from their office desk, Starbucks or even Bangladesh nothing changes as far as trust goes. There are no if/then rules for how strong the authentication is, based on whether the user is internal or external. As far as the app is concerned, all users are external. We implement MFA across the board, every user, every time, everywhere. We get smarter about authentication and ask a lot more questions of the user by including behavior patterns, and a more in-depth look at the user's attributes. Does the user know the password and are they logging in from a device we've seen before with an IP address we've seen before and from a location we expect? That’s a very different set of questions from a simple user authentication process that asks, “Does the user know the password and some random number?”
The second concept is that we question everything. We’ve verified who you are; now we need to verify what you are doing or asking to do. Are you trying to access payroll data during business hours from an IP address we see often? Including more context into our authorization decisions allows us to make access decisions based off of what’s happening right now instead of some static rule we set 6 months ago. We need the ability to build dynamic authorization policies that include information about the user and about what the user is trying to do. We need to add behavior patterns to understand what activity is considered normal and which are abnormal. Zero Trust means we have to be smarter and more in-depth in the actions we take to verify a user.
Everything we’ve talked about all centers around identity. Having a robust identity infrastructure gives you the ability to build more dynamic and identity-aware applications. Administration and governance is the first brick that must be laid to develop a robust solution. It’s your foundation. A trusted source is required to pull attributes needed for authorization policy and deeper authentication. The process of feeding that source and ensuring that the attributes are accurate is the where governance and administration come into play. This process allows you to control the lifecycle of new accounts and attributes and ensure they are aligned both from a quality perspective and policy perspective.
These concepts combine to give you a type of "Zero Trust" architecture. However, I want to take the time to make a crucial point about Zero Trust: it's not a solution. Zero Trust is a way of thinking, or perhaps even better stated, an approach. The entire concept is to challenge you to think differently about how you build your applications, networks, and security controls. You begin with the statement that you don't trust any user. You don't depend on a single attribute to determine your level of trust; instead, you continually build that trust with the user by asking questions. Who are you, where are you coming from, what are you trying to do, when are you trying to do it, etc., etc.? The common component for all of these questions is that you are trying to establish the person's identity so the more you establish their identity, the more you trust them, the more you trust them, the more access you give them. Identity remains central to good security posture, and in this Zero Trust world that doesn't change, in fact, it enhances the need for it.