Partner TechTip: A Decade of IBM i Security: The Good, the Bad, and the Downright Ugly

Security - Other
Typography
  • Smaller Small Medium Big Bigger
  • Default Helvetica Segoe Georgia Times

PowerTech leaks details from its 10th "State of IBM i Security" study.

 

A decade has passed since PowerTech published the first hugely popular "State of IBM i Security" study; however, despite the evolution of the server and the operating system, many configuration vulnerabilities remain the same.

The Good…

Each year, PowerTech audits hundreds of servers using a proprietary assessment tool. Clients have the option of sharing statistical information—devoid of any application or identifying data—for inclusion in the study. The statistics are culled from different servers each year, which makes year-to-year comparisons difficult. But the volume of assessments and the variety of server and business sizes demonstrate legitimate insight into how the majority of customers approach IBM i security.

 

While there are more statistics than we can cover in a short article, here's a peek from the six areas of the latest study, which included:

  • 101 servers
  • 1,082 users per server (average)
  • 427 libraries per server (average)

 

An analysis found all the servers running on supported versions of the OS. Although IBM declared support will end in September, 41% remain on V5R4. Another 41% were on V6.1, and the remaining 17% were running IBM's latest: V7.1.

 

The study reports ongoing challenges in the deployment and configuration of security controls, but the best news is none of these issues represents vulnerabilities in IBM i itself.

The Bad…

Security best-practices start with the configuration of numerous system values. Although often undermined by other settings, the system security level (QSECURTY) sets the tone. Well-documented vulnerabilities in security level 30 and below led IBM to establish level 40 as the recommended minimum. 37% of servers were operating below this level. Numerous other values, including Allow Object Restore (QALWOBJRST) and Verify Object on Restore (QVFYOBJRST) remained at their shipped values, belying a typical configuration of "load and go."

 

Most IBM i environments suffer from overly privileged users. An average of 65 users operate with All Object (*ALLOBJ) special authority, providing unrestricted access to every object—including other users' profiles! 123 users carry Job Control (*JOBCTL), allowing the system to be brought down to a restricted state. These types of capabilities should be reserved for administrators.

 

IBM i contains a comprehensive auditing facility, but 12% were not utilizing it. This means zero visibility to critical system events like unauthorized access attempts, data files being saved or deleted, and changes to system values. Despite being at its lowest level in recent years, many organizations were not auditing all of the recommended events, and only 27% had a recognizable audit reporting tool installed.

… And the Downright Ugly!

No one wants to be the "grim reaper" of IBM i security, but after a decade of seeing severe and persistent issues, it's difficult to report a positive spin on the next statistics.

 

Profiles proved to be a significant source of vulnerability. Here are some averages to consider: 79 users had default passwords, and half of those were enabled and ready for use! 197 enabled users had command-line permissions. 31 servers never required users to change their passwords. 9 servers permitted an unlimited number of attempts to sign on, although you'll have to download the full study to find out the staggering number of invalid sign-on attempts we found.

 

Unlike most operating systems, where no authority means no access, IBM i provides a default level of authority called *PUBLIC. We found that only 8% of libraries had *PUBLIC authority set to the ideal level of *EXCLUDE.

 

Many organizations rely on legacy user restrictions such as menus, application security, and command-line restrictions. However, modern interfaces such as FTP and ODBC provide direct access to the database. IBM i supports exit programs to supplement native object security, but 69% of servers remain wholly unprotected by even a single exit program.

 

Unfortunately, many of these statistics haven't shown consistent signs of improvement. Despite the fact that some of the most critical enterprise data is found on these servers, the overwhelming evidence of weak configuration supports the conclusion that companies are still not giving the necessary attention to IBM i security.

Does the Story Have a Happy Ending?

Yes! However, it's critical that organizations recognize that security for the server does not come preconfigured. It doesn't have to be overwhelming; many easy-to-implement changes can reduce risk.

 

It's up to you to ensure that your server meets regulatory or legislative standards, or simply best practices. Use this awareness to plan to review and remediate your security vulnerabilities…before you become a statistic.

Next Steps

My suggested call-to-action includes:

 

  • Download the complete 2013 "State of IBM i Security" study at www.powertech.com.
  • Request a free Compliance Assessment to determine the level of risk inherent to your own server configuration.
  • Increase your knowledge of IBM i security. PowerTech has a growing number of white papers and articles available on our website.

 

 

BLOG COMMENTS POWERED BY DISQUS