Is securing your network worth the money?

Security researchers detail P2P threats, vulnerability disclosures and hacker profiling at recent Carnegie Mellon confab.

Current Job Listings

P2P threats, vulnerability disclosures and hacker profiling are among topics being researched by experts in economics and IT security.

We often hear corporate IT pros complain that justifying security expenses is tough because they don’t necessarily generate revenue or enable new business opportunities. In fact, figuring out the economics of IT security is so challenging for customers and vendors that lots of the world’s best researchers are putting their minds to the task. They recently shared results at a conference hosted by Carnegie Mellon University. Here’s a sampling of the research presented:

“Inadvertent Disclosure – Information Leaks in the Extended Enterprise”

A pair of researchers from Dartmouth College’s Center for Digital Strategies (Tuck School of Business) examined involuntary leakage of information via peer-to-peer file sharing networks at a group of large financial institutions (see the paper). What they found wasn’t pretty.

Using technology from Tiversa to gather and categorize tens of thousands of P2P searches and files related to the top 30 U.S. banks, the researchers found “significant information risk firms and individuals face from P2P file sharing networks” such as Gnutella, FastTrack and edonkey. They found that end users are often lazy or poorly organized, leading to internal company documents as well as partners’ documents being revealed. “For one bank, we found a spreadsheet with 23,000 business accounts including their contact names and addresses, account numbers, company positions, and relationship managers at the bank,” the report states. Also found: one bank’s security review process manual.

The researchers also addressed the issue of “digital wind,” which entails getting your company searched on because your company or brand names are similar or the same as those of things such as bands or songs (State Street, for example, could get caught up in searches for the Death Cab for Cutie song “State Street Residential”).

The researchers found that criminals actively search P2P networks looking for information to exploit.

While some companies and ISPs use port filtering to block or throttle P2P traffic, many users of P2P systems have moved to new systems and some of the newer systems have been designed to use ports also used by Web, e-mail and other traffic.

The researchers’ recommendations include introducing “file naming conventions and policies to reduce the metadata footprint of their documents.” Employee/partner education about P2P dangers is also key, in combination with technical solutions, they wrote. Benchmarking can be done to measure document leaks per employee over a set period of time.

One other conclusion: Big firms have to work a lot harder to safeguard themselves against P2P-oriented leaks than smaller ones.

“Network Security: Vulnerabilities and Disclosure Policy”

A pair of Tel Aviv University researchers and another from the Michigan State University Department of Economics examined “the economic incentives facing software vendors and users when software is subject to vulnerabilities” in their paper.

The researchers lay out the dilemma faced by software vendors who don’t want to risk harm to their reputation by disclosing vulnerabilities but also don’t want ill-intentioned hackers to exploit vulnerabilities that they stumble across. Also, the paper notes: “the firm’s disclosure policy and its profit-maximizing behavior are interdependent.” The researchers develop a model that looks at the ramifications for software vendors when they disclose, don’t disclose and put more resources into product security ahead of time.

Among findings is that “bug bounty” programs offered by some software makers can boost their profitability, as long as they actually disclose the vulnerabilities. Logically, only those customers who actually install updates benefit from the programs.

“The Countervailing Incentive of Restricted Patch Distribution: Economic and Policy Implications”

Three Purdue researchers took a look at what difference it makes to software vendors and customers, as well as the government, if a software vendor restricts distribution of patches to only those with legal copies of its software (such as with Microsoft’s Windows Genuine Program). Issues include whether restricting the patches actually does more harm than good in that this leaves lots of vulnerable software implementations out there.

The researchers conclude in their paper that: “If the cost of the patch quality is high, then in the presence of intense anti-piracy action the vendor does not benefit from vertically differentiating between the legal and the pirated versions.” Also: “If the population is less likely to pirate, intense anti-piracy enforcement should be complemented with the decision to restrict the patch. However, we notice that there exists a level of anti-piracy effort by the government above which vertically differentiating the legitimate copy from the pirated copy is not optimal.”

“Optimally securing interconnected information systems and assets”

A threesome of Carnegie Mellon University researchers address this long and involved question: “How should a multi-divisional enterprise optimally deploy security countermeasures in the context of heterogeneous information systems possessed by its divisions and in response to different kinds of damages that threats can cause to the enterprise’s information systems and assets?”

The upshot of the work is that “there are strategic issues in information security decision making and that the distortion due to incomplete knowledge of information systems by the CIO has to be weighed against incentive problems when division managers make decisions.” The team looks at threats to network availability and confidentiality (personal info), and finds that dealing with availability losses is easier “since the deployment decisions involving protection countermeasures at each division are strategically independent.”

“Interdependence of Reliability and Security”

This paper studies incentives for vendors to invest in solidifying the reliability and security of their software. The authors, from the University of Michigan, University of California at Berkeley and HEC Montreal, find that users typically can’t see the difference between whether a failure happens because of a security or reliability shortcoming and trace “reliability and security failures to the same source – software bugs.”

“Measuring Security Investment Benefit for Off the Shelf Software Systems – A Stakeholder Value Driven Approach”

USC researchers presented a method for quantifying security threats that is designed to help organizations determine how much to spend on off-the-shelf security technologies to address their security needs. Their paper presents a case study of one large organization using the methodology. “Compared to traditional security risk management approaches,” the authors write, “T-MAP (Threat Modeling method based on Attacking Path Analysis) does not require accurate values of probabilities, frequencies and size of loss which are usually very difficult to estimate in practice.”

“Cyber-Insurance: Copula Pricing Framework and Implications for Risk Management”

This study by University of Buffalo and Brock University researchers looks at pricing for insurance policies to cover computer security breaches in light of massive losses cited by organizations that have been affected. The researchers develop a model for pricing cyber-insurance but acknowledge doing so had its limitations in that there is limited data on cybercrimes to work with. The researchers urge IT managers to collect good data on security breaches to help themselves when it comes to negotiating insurance premiums.

“Mental Models of Computer Security Risks”

Indiana University researchers explore the difference in perceptions of security risk by experts and non-experts, and how organizations can benefit by communicating with end users about risk based on “mental models” used by the non-experts. (See their paper.)

“Economics of User Segmentation, Profiling and Detection in Security”

The pros and cons of profiling users to ferret out hackers is explored in this research out of the University of Texas at Dallas and Northwestern State University in Louisiana. Among the issues: Hackers faking their profiles to skirt the system. However, they argue that profiling can be beneficial, even if some legitimate users need to jump through a few extra hoops now and then.

For more on network research, read our Alpha Doggs blog.

Learn more about this topic

15 bleeding-edge network research projects you should know about

Alpha Doggs network research blog

Illegal P2P use dropping

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.
Now read: Getting grounded in IoT