Understanding Patch Management

It’s like a recall on your car…

A Jolley | Hadaway Article.

If you like to stay current on technology news, you’ve probably found it hard to ignore the recent stories about a major security problem with Intel processors, and the rush from software and hardware companies to mitigate the issue.  Most of the solutions to this problem come in the form of a patch, in many cases downloaded and installed by the operating system automatically. Given that patch management is such a time consuming, confusing, complex, and often highly visible (when not handled right) control, management should understand what is meant by the terms “Patch Management” and “Vulnerability Management.”

If you have already read our R-7 article, you know that we at infotex believe that Vulnerability Management is going to be one of the most critical controls to circle back around to in 2018.  So much that we have two versions of this article.  (We have a technical version of this article here.)  Yes, we believe you should start educating management on the basics of Vulnerability Management, which we consider a super-set of Patch Management.  We are now proclaiming, in one of Dan’s manifestos, a belief that Vulnerability Management = Policy (board approved, management educated)+ Awareness + Prioritization + Patch Management + Testing + Tweaking.

Awareness, you ask?

Yes, awareness! And not just technical awareness. We believe management should be brought on board what we’re doing, if for no other reason, in order to help soften the blow when we fail. Failure usually comes in the form of an audit test, but if the failure is a result of a breach, management will be far more cooperative if they understand Vulnerability Management. We believe this is the missing puzzle piece for those institutions still struggling against the patch management findings in their audit reports.

To teach nontechnical management, we use metaphors. And for this topic, we’re choosing the auto industry as the metaphor.  Like the auto industry, people design software, and they later learn there was a mistake in the design.  If a design flaw is discovered in the auto industry, the manufacturer will issue a recall or send instructions on how auto owners can repair the design flaw.

Like the auto industry, application publishers register purchasers to notify them of design flaws as they are discovered.  Application publishers release updates to fix their software mistakes.  We call these updates “patches.”

In application design, most flaws are related to the use of the application.  For example, a report might not print right, so a patch would fix that process.  Some flaws, however, open security holes in the application that can be exploited by malicious persons.

The problem is, in application design, products are released with thousands of flaws, because time to market is prioritized over usefulness and even security. If the application used to read this article was a car, it would have NEVER been published. As a result, the volume of patching that users wrestle with daily is extraordinary, requiring a method of prioritization.

We recommend smaller institutions simplify the prioritization process by sorting patches between critical and non-critical patches.  If the flaw causes a security vulnerability, the patch that fixes the flaw is considered a critical patch.   Critical patches should be applied immediately.  The “patch management systems” used by your IT staff identify which patches are critical and which are not, which helps them prioritize the patch management process.

Complicating matters further, sometimes patches will cause software failures.  Imagine if your Chevy had active parts in it built and controlled by Ford, Volkswagen, Chrysler, and even AMC, which no longer manufactures cars.  To fix one part, you take a chance that the other parts will no longer work with the part being fixed.

This phenomenon is rampant in the application world.  We can never know for sure if a patch will cause failures, and thus patches should be tested prior to deployment.  The easiest way to test a patch is to install the patch on one computer and use that computer until we’re sure the patch didn’t cause issues for that computer.

Testing patches this way is often very cumbersome and takes a lot of valuable time. Thus, for noncritical patches, most small institutions just wait to see if there are any issues discovered by larger institutions.  In other words, they delay application of patches for a few weeks to see if problems are reported.

This approach does not work for critical patches, however, because the bad guys are aware of the vulnerability, and have ways to find the networks with those vulnerabilities.  Thus, critical patches must be tested as soon as published, so they can be applied as vulnerabilities are discovered.  For smaller community banks, critical patches should be applied within 24 hours of their availability.  The longer it takes to apply a critical patch, the more exposure the bank faces to security incidents.

Finally, we pay close attention to the performance of our auto just after the dealer has fixed the design flaw, because we know that if it wasn’t done right we want to catch it while we can still take it back to the dealer.  Likewise, in application management, our technical team needs to test systems to ensure that applications patches have been properly and thoroughly applied.  Patch management systems are so complex, that tweaking them out requires a test-feedback-tweaking process until we are satisfied we are appropriately executing our patch management policies.  Most smaller institutions are now investing in Nessus or other software, so they can do their own vulnerability scans on a periodic basis for this very purpose.

Management usually only hears the word, “patch,” during audit reports, when our processes have failed to cover all critical patches.  By understanding the vulnerability management process, management will be able to help the technical team overcome the many obstacles to a smooth operating vulnerability management process.

We hope that sharing part or all of this article with your management team will help with your vulnerability management processes. Let us know if you have anything to add, or disagree!

Check out the technical version of this article here.

Original article by Dan Hadaway and Matt Jolley.