Pen Testing Checklist for Cloud

Start with the Contract. Your Cloud services are provided under contract between you and your CSP. This forms the base of the relationship, and defines what activities each party is responsible to perform. Not all CSP’s are the same, nor are all contracts identical. Some will have various tiers of service, others may provide a base offering with additional “add-on” options. Whatever your situation, it is vital to have a clear understanding of R&R, policies, service commitments, and restrictions.

1. Check the Service Level Agreement (SLA) to ensure the appropriate Pen Test policy has been identified, and R&R clearly defined. In many cases, elements of Pen Testing are spread across multiple players such as the CSP and the client, so it is necessary to clearly document who does what, and when it is to be done.
2. Governance & Compliance requirements need to be understood. Factors need to include which party will be responsible to define, configure and validate security settings required to meet applicable regulatory controls for your business. This includes providing appropriate evidence for audits and inspections.
3. Security and Vulnerability Patching and general maintenance responsibilities and timeframes need to be documented. You as the client may have responsibility for maintaining your virtual images and resources, but the CSP will likely be accountable for the underlying physical hardware systems. Both need to be actively managed, along with all network and SAN equipment.
4. Computer access and Internet usage policies need to be clearly defined and properly implemented to ensure appropriate traffic is permitted while inappropriate traffic is denied at the perimeter.
5. Ensure all unused ports are disabled and unused protocols are either not installed or disabled and locked down to prevent unauthorized activation.
6. Data encryption while both in transit and at rest is becoming more common, but never assume. Ensure that encryption is either set as the default or that appropriate steps are implemented to ensure it is activated.
7. Verify that your requirements for Two Factor Authentication and One Time Passwords are implemented and actively securing network access. Check if your CSP permits any bypass scenarios.
8. SSL is only as good as the Certificate Authority (CA) that issued the certificates. Ensure SSL is active, and that a reputable CA stands behind the certificates.
9. Hold your CSP accountable and validate they are using appropriate security controls for physical and logical access to the data center and the infrastructure hardware with which they provide your services.
10. Know your CSP’s policy and procedures relative to data disclosure to third parties, both for unauthorized access and providing data when requested or subpoenaed by law enforcement.

More than just running a scan, Pen Testing requires an understanding your environment and the associated roles and responsibilities and even liabilities between you and your CSP.

PCI Pen Testing Requirements

Scanning Isn’t Testing

Requirement 11 is the juggernaut of PCI v3.2. It’s packed full of objectives involving network scanning (11.2) and penetration testing (11.3). Just because scanning and testing are lumped into the same PCI requirement, doesn’t mean they accomplish the same goals.

The PCI Standards Security Council is helping to clear the air with a couple of simple descriptions.
Vulnerability Scanning: Identifies, ranks, and reports on vulnerabilities that, if
exploited, may result in an intentional or unintentional compromise of a system. Translation? It’s an automated process taking a short amount of time, with no verification and very high chance of false-positives.
Penetration Testing: Identifies ways to exploit vulnerabilities to circumvent or defeat the security features of system components. Translation? It’s a manual or automated testing process that uses vulnerability scanning results as a baseline, lasting days or weeks depending on the scope.

Testing Ain’t Easy

Whether you use a tool or a service, complying with PCI Requirement 11.3 is a formidable task taking a combination of resources, time and a little bit of planning. In its simplest form, Requirement 11.3 mandates internal and external penetration testing at least annually or after “any significant infrastructure or application upgrade or modification”. Considering the volume of iterations of common operating systems and applications, I’d say we can throw out the idea of only testing annually. So, how do you test regularly and stay agile?

1. The right person or service: Take a look to see if any of your team members have pen testing experience. The great thing is that some penetration testing products, like Core Impact, have the option of using wizards to run various tests. This helps automate some of the more laborious tasks, while accomplishing the same goals in less time. As a supplement to this, try bringing on a more seasoned penetration tester or dedicated third-party security service.
2. The right tool: Depending on your needs and overall requirements scope, the tool you require may differ. Whatever it is, you need to be able to test regularly as things change in your environment. So, if you have one person on your team who only uses an open-source command-line based pen testing tool, with no other options, this could leave you in a bit of a pickle if he or she leaves. Be sure to have a backup plan.
3. The right methodology: Penetration testing doesn’t just happen and disappear. According to the new PCI Security Standards Council penetration testing guidelines, it requires a phased-approach involving pre-engagement, engagement, and post-engagement steps. These steps cover important topics such as timing, frequency, reporting, systems to target, success criteria etc. Preparing properly will help you stay on target, meet the PCI requirements, and not become overwhelmed. Well, not too overwhelmed at least.