Good Software Hygiene is Effective in Combat of Malware-Driven Data Breaches

Wolfgang Kandek

Last updated on: September 6, 2020

On Friday April 15th, The Oak Ridge National Laboratory (ORNL) disconnected its Internet access to contain an intrusion and interrupt the theft of data. Attackers had gained access to the ORNL network on April 7 through a phishing e-mail attack carrying malware with an exploit for a 0-day vulnerability in Microsoft Windows Internet Explorer.

Previously, we had seen a similar attack on the security company RSA, where data related to SecurID, RSA’s two-factor token authentication product was extracted. In RSA’s case, the phishing e-mail involved an Excel spreadsheet purporting to be about the hiring budget for 2011. The spreadsheet contained an exploit for a 0-day vulnerability in Adobe Flash.

At the same time Verizon’s 2011 Data Breach Investigations Report (DBIR) affirms for the 3rd year in a row that the majority of data breaches (96 %) could have been avoided with the implementation of simple countermeasures.

Organizations can effectively protect themselves by implementing good software hygiene, which starts by introducing a structured patching process aimed at installing critical updates for all software within a short timeframe, we recommend within 10 days. Organizations that have implemented such fast patching have seen a significant improvement in the robustness of their infrastructures and have been documenting their progress publicly (see reference section on processes in use at Goldman Sachs and US State Dept).

Fig 1: Motivation for Patch Speed at Goldman Sachs (From SPO-208 RSA US 2009)

Fast patching will prevent infection from all of the common malware exploit kits that are available for purchase. The toolkit “Phoenix 2.5” for example offers 5 exploits based on the PDF file format, 3 on Java and 1 each for Quicktime and Adobe Flash, all of them abusing vulnerabilities that are already patched.

Further resilience can be gained by controlling installed software and its configuration. The ORNL case would have been countered by the consistent use of an alternative browser. The Excel attack could have been prevented by prohibiting active content in Microsoft Office Trust Center or uninstalling Adobe Flash, preferably both. Switching to a more modern version of the base OS or even an alternative OS will also help to add resilience against malware (i.e. Windows7 64bit, Mac OS X or Linux).

This level of tightening of IT configurations raises the bar significantly and will keep most classes of attackers out of enterprises networks. Talk to your industry peers to see what they are doing; a number of organizations are already operating their networks in this way and can attest to the effectiveness of these measures.


Show Comments (2)


Your email address will not be published. Required fields are marked *

  1. Great article. It is difficult to roll out patches so quickly though in a large environment with 100’s of both Vendor & Home Grown applications that all depend on different versions of plugin’s and other 3rd Party applications. The testing & analysis of patches' impact against all these applications on various desktop/server builds alone can take weeks if not months. It’s often an uphill battle every time…

    1. Good to hear you liked it.

      I believe that extensive patch testing can be very time consuming. Goldman Sachs dealt with the same issue, but ultimately decided it was better to have a robust infrastructure rather than delay the rollouts because of some outliers and put the whole company at risk. They do provide for such instances with an exception and usually try to get some other security measure implemented, for example prohibit Internet/E-mail access for such a non-standard machine.

      What has your experience been ? Have you seen impacts from patches ? Has it been worth doing the drawn out, slow testing ?