PCI Series: Requirement 1 - Install and Maintain a Firewall Configuration to Protect Cardholder Data

This article is one of the blog posts dedicated to use of Identity Management (IdM) and related technologies to address the Payment Card Industry Data Security Standard (PCI DSS). This specific post is related to requirement one - install and maintain a firewall configuration to protect cardholder data. The outline and mapping of individual articles to the requirements can be found in the overarching post that started the series.

The first requirement of the PCI standard talks about the firewalls and networking. While Red Hat's Identity Management solution is not directly related to setting up networks and firewall rules, there are several aspects of IdM that

need to be mentioned in this context. The first is that IdM servers can be deployed inside and outside a firewall. In either case IdM servers need to communicate with clients and to each other using the LDAP and Kerberos protocols.

IdM servers that are deployed inside the firewall create challenges for authenticating clients that are located outside the firewall on a separate network or in a DMZ. The IdM solution leverages Kerberos heavily. The main reason for this is that the Kerberos protocol ensures that end user passwords are not sent "over the wire" thereby reducing the risk of password interception or leak. However the use of Kerberos creates a challenge for administrators who traditionally had to open a Kerberos port in the firewall to allow the authentication to go through. This, in many cases, is a non-starter. The IdM version that comes with Red Hat Enterprise Linux 7.2 includes a feature called KDC proxy. Several years ago Microsoft authored a standard that allows for proxying the Kerberos protocol over HTTPS. KDC proxy is the open source implementation of this protocol. This solution avoids the need to open a Kerberos port in the firewall and leads to a tighter firewall configuration that is in the spirit of the PCI DSS standard.

The solution still requires opening an LDAP port so that clients can download identity information. For purposes of identity lookup the IdM server in the DMZ can act as a proxy between clients in the DMZ and Active Directory (AD) servers behind the firewall. The firewall rule in this case can be set to allow connection only from the IdM server host in the DMZ to AD inside the firewall thus significantly limiting the attack surface. Placing an IdM server in the DMZ to serve clients there enables a more secure integration of those systems into an AD fabric.

The other aspect that is worth mentioning is IPSec VPNs. The IPSec VPN specification has been extended to allow for Kerberos authentication. The implementation of IPSec VPN (libreswan) is underway. This enhancement combined with placing IdM outside the firewall will allow a VPN user to authenticate against an IdM server first using, for example, OTP authentication over Kerberos, to then acquire proof of authentication (ticket), and (finally) to connect to the VPN server without being prompted. Such an approach, when integrated with desktop login, would allow for signing into the network and logging into the system at the same time - eliminating multiple steps and prompts.

Questions about how Identity Management relates to requirement one?  Reach out using the comments section (below).