Regardless of whether it's for PCI, HIPAA, SOX, or GLBA, chances are high that if an auditor's bound for your organization, your access control is about to go under the microscope. With so many compliance-driven mandates around separation of duties and user monitoring dependent on strong access control regimes, it's no wonder that this is one of the key areas that auditors will focus on. After general policies, standards, and procedures, access control is one of the first things we examine in our audits. The following are some key access control considerations during an assessment.
1. Multidimensional Authorization and Authentication
Before delving into access control specifically, you should stop and imagine a bull's eye with access to the data itself in the center and concentric circles around that. Each subsequent circle stands for the application, database, operating system, network, and then finally the physical data center layers. An interruption or compromise of any of the outer layers will affect access to those inside. What are the authorization controls and what are the authentication controls for every layer of that bull's eye? Authorization is user account [approval] to get access to their data, and then authentication controls are going to be password controls. On both counts, we'll be looking for policies, standards, and procedures for ensuring security and reliability of access and user monitoring. So for authorization, that includes things as simple as policies around naming conventions, as well as definitions of what accounts can do on the system and user approval processes. For authentication, these policies include how often passwords are reset (commonly between 30 to 60 days) and length of passwords, with seven being the safe bet in the industry at the moment. Additionally, we'll also be looking at how passwords are transmitted and stored. The minimum security we're going to have around the passwords is that it has got to be encrypted – encrypted at the time of entering and encrypted during storage. Not only does this prevent malicious hackers from stumbling on a treasure trove of passwords, but it also is critical for maintaining the integrity of separation of duties.
2. Role-Based Access Control
Another important aspect to enterprise access control that's on our hot list is that the organization engages in role-based access control (RBAC). Ideally, organizations should be streamlining the authorization process by creating user roles based on the rule of least privilege. Well-defined roles and good administration of the provisioning and de-provisioning to roles make it easier to ensure users have access only to what their job requires. And every time a new user comes into the organization, we don't have to add in his or her user ID to that particular set of database resources; we add them to the role. That role is the one that maintains the access and as such, it makes it a whole lot more efficient for adding users to the system. Then their access is going to be controlled based on their job responsibility.
3. Privileged User Protections
In order to keep application and database environments humming, there's naturally going to be some IT super-users out there who will be able to touch more data than the common worker bee. Those are people that you should be logging everything they do. Typically, there's an administrator account that has access to the database that can do anything and everything. You want to monitor that account. But we’re not just looking for straight logging. We’re also looking to see whether the organization has instituted a way for specific users to log in to that root account through some sort of intermediary account that will allow for monitoring not just by the account, but by the user controlling that root account at any given time.
4. Filtering and Parsing User Activity
Effective access control doesn't just put up barriers to entry. It also enables for more effective and granular visibility into what specific employees are doing within specific systems. One of the biggest problems we see is around the compliance logging required by many regulations to monitor user activity. Often that log data remains untapped, with organizations failing to use a filtering mechanism or log management tool to parse data. You're interested only in those logs that meet a predefined set of parameters that are of concern to the organization. Say a password after five attempts is automatically suspended and then you see another attempt tried five times and that's also suspended. If the parsing tool can interrogate all that data, that's the alert we want sent to the security administrator or the analyst to investigate or follow. PCI is very black and white about having daily log reviews. If you have high alerts come in, what are your processes to deal with them, for example.