Original URL: http://www.theregister.co.uk/2005/11/09/automatic_graylisting/

Automatic graylisting of unwanted software

Maximum security, minimal effort

By Todd Brennan

Posted in Security, 9th November 2005 14:09 GMT

As network perimeters become ever more porous, and endpoint security becomes even more critical, companies today are struggling with the problem of unwanted software - whether it's new, unknown, and potentially malicious software, or simply known but non-business applications.

Now, a new approach to endpoint security known as automatic graylisting is enabling IT professionals to regain control over spyware, malware, and other unwanted and unapproved applications with real-time, network-wide visibility and control - for maximum security with minimal effort.

The problem of unwanted software

Unwanted software represents a root cause of a wide variety of IT problems that involve not just time, but money. For example, Microsoft estimates that spyware causes one third of system crashesand according to the Business Software Alliance, unlicensed software can result in settlements in excess of $100,000.

Unwanted software includes worms, viruses, spyware, vulnerable applications, unlicensed software, and unsanctioned applications, and it can arrive on corporate networks in many ways. The scope of the problem ranges from mobile users connecting to infected networks to end users unknowingly infesting their office desktops with spyware. As well, end users can knowingly download non-sanctioned applications such as peer-to-peer file-sharing, instant messaging, mp3 applications, and others.

SANS has reported that the number of vulnerabilities in 2005 is going up, not down: "Despite increasing public and corporate awareness about cybersecurity, the number of computer vulnerabilities in the second quarter of 2005 increased 10.8 per cent compared with the first quarter, according to a new survey from the SANS Institute, which develops data and research on information security. In all, SANS discovered 422 new vulnerabilities, up from 381 in the first quarter."

To complicate matters further, spyware producers continue to make the legal case that their software is neither malicious nor illegal. While unwanted software encompasses a variety of applications that have different implications and characteristics, today's solutions typically focus on narrow aspects of the problem. Thus, different tools have evolved to solve a subset of the problem. For example, most current security products attempt to detect only malicious applications of certain types and new malicious software is discovered each day. In fact, Sophos recently identified 1,233 new viruses in September 2005 alone. Unfortunately, this approach has caused a proliferation of agents on systems that increase complexity and administrative effort.

According to Gartner's Security Software Forecast on 30 March, 2005, "revenue for the security software sector will grow at a compound annual growth rate of 16.2 per cent through 2009, with new license revenue reaching $11.4bn." This double-digit growth indicates that current solutions to today's problems are simply not working.

Today's solutions

Today's solutions to the problem of unwanted software take a narrow view and typically target just one problem, such as spyware. This requires IT administrators to install and manage an increasing number of agents on their systems. Furthermore, most products depend on perpetually out-of-date "black list" technology that compares executable files against a "known bad" list of bit patterns or behavior signatures. This approach leaves systems exposed to new, unknown, and potentially unwanted software.

Other products use a "white list" approach that limits systems to a "known good" list of executable files. While this approach provides the best protection in theory, these solutions are often overly restrictive and the administration of white lists is daunting.

When these approaches fail, the result is late detection, false alarms, and complexity. Furthermore, none of these approaches gives IT visibility and control over all the executables in their infrastructure. Clearly, what is needed is a solution that enables administrators to see and monitor what has arrived, when it has arrived, who has executed it, and where it is now.

Introducing automatic graylists

So far what we have learned is that black lists are eternally incomplete and white lists are difficult to maintain in dynamic environments. In desktop environments, new applications, plug-ins, and Active-X components - wanted or unwanted - appear frequently and do not yet exist in a black or white list. Both approaches require additional information to increase security coverage and reduce administrative burden.

Now there is a powerful new mechanism to control unwanted software that integrates with existing IT processes, called an automatic graylist. Unlike other alternatives that attempt to detect what's wrong or bad, automatic graylists track new unknown files - before they have been classified - and help automate their approval. When combined with white lists and black lists, automatic graylists overcome the drawbacks of today's other endpoint security solutions while providing the best possible protection.

Automatic graylists do not rely on malware signatures or behavioral patterns and allow the customer, not the vendor, to decide what software is appropriate and approved to run on their infrastructure.

In this new model, new application files are detected in real time as soon they appear on systems and are automatically added to the automatic graylist. They can be easily approved or banned, based on current security policy. This supports a dynamic environment, allowing the administrator to define the appropriate security policy for groups of desktops, laptops, and servers. With this information, IT professionals can create policies that automatically track and control software that has not been centrally rolled out or pre-approved by IT. When applied on a network-wide basis, this automatic graylist approach provides the best protection against unwanted software without a complex or costly administrative burden.

For host groups where graylisted software is blocked, this approach provides zero-day protection as new, unknown files are effectively put on probation. For example, enterprise customers could block new files for a period of time, e.g. 30 days, after which time the files are dropped from the automatic graylist. This buys customers time to react while anti-virus signatures catch up with new zero-day attacks.

Be prepared

IT managers, security managers, and executives are watching with mounting concern as unwanted software on enterprise endpoints increases. Furthermore, current security products offer, at best, limited protection against these applications and the costly damage associated with them. Current technologies do not provide enough early detection and lockdown capabilities to deal with new unknown and potentially malicious code, which is typically the most disruptive. And enterprises cannot respond quickly enough to avoid harm.

Now, with the approach of automatic graylists, security administrators for the first time ever have gained visibility and control over unwanted files that install and execute on user systems.

Copyright © 2005, SecurityFocus

Dr. Todd Brennan is Chief Technology Officer of Bit9, a security company that focuses on endpoint systems. Dr. Brennan also founded Okena, which dealt with emerging computer threats and was acquired by Cisco in 2003. He has been a research staff member at MIT Lincoln Laboratory, and he holds Ph.D. and M.S. degrees in Electrical and Computer Engineering from the University of Wisconsin.