Define the surface to be protected, map transactions, architect the network, create policy and then implement monitoring of traffic, accounts and data — those are key steps to a ‘zero trust’ strategy even in Active Directory environments, according to Lepide.
Aidan Simister, writing for the data security platform provider, said that developing full understanding of who and what should have access to a resource, its location and how or why it is used, remains important.
“You must ensure that you are monitoring all events relating to the way your user accounts and sensitive data are being accessed and used,” Simister said.
The Microsoft Active Directory environment that stores information about objects on the network and exposes this information for administrators and users qualifies as a critical resource which must be protected, he says.
That means looking at solutions that can scan repositories wherever they are and automatically classify data according to a chosen schema. And never mind passwords — ‘zero trust‘ means not even administrators and admin-level privileges can be ignored.
“When an admin account requires elevated privileges, for whatever reason, they must be granted what is called just-in-time access, where access rights are temporarily granted and then revoked as soon as they are no longer required,” Simister said.
Simister suggested adopting a multi-factor authentication solution that integrates with Active Directory and warns users not to automatically authenticate on-prem users with Azure.
“The policy is to never trust, always verify. As such, you would be better off asking your employees authenticate themselves each time they need access to a critical resource,” Simister said.
That also meant monitoring all access to sensitive data of any kind. While it is theoretically possible to manually scrutinise server logs for suspicious activity, this is likely a slow, painful and error-prone process.
“A better approach would be to adopt a dedicated real-time auditing solution which will display a summary of important events via a single dashboard,” Simister said.
Most sophisticated offerings today use machine learning models to automatically detect and respond to anomalous user activity with real-time alerts sent to an inbox or mobile device, he added.
“Modern IT environments are a lot different from 15 or so years ago. They have become more distributed and dynamic, with employees accessing their corporate network from various locations and devices,” Simister pointed out.
( Photo by Kristin Wilson on Unsplash )
