On Computing Enterprise IT Risk Metrics

External Posting Date: February 21, 2011 [Fulltext] Approved for External Publication Internal Posting Date: February 21, 2011 [Fulltext]  On Computing Enterprise IT Risk Metrics Sandeep Bhatt, William Horne, Prasad Rao HP Laboratories HPL-2011-26 Assessing the vulnerability of large heterogeneous systems is crucial to IT operational decisions such as prioritizing the deployment of security patches and enhanced monitoring. These assessments are based on various criteria, including (i) the NIST National Vulnerability Database which reports tens of thousands of vulnerabilities on individual components, with several thousand added every year, and (ii) the specifics of the enterprise IT infrastructure which includes many components. Defining and computing appropriate vulnerability metrics to support decision making remains a challenge. Currently, several IT organizations make use of the CVSS metrics that score vulnerabilities on individual components. CVSS does allow for environmental metrics, which are meant to capture the connectivity among the components; unfortunately, within Section 2.3 of [1] there are no guidelines for how these should be defined and, consequently, environmental metrics are rarely defined and used. We present a systematic approach to quantify and automatically compute the risk profile of an enterprise from information about individual vulnerabilities contained in CVSS scores. The metric we propose can be used as the CVSS environmental score. Our metric can be applied to the problem of prioritizing patches, customized to the connectivity of an enterprise. It can also be used to prioritize vulnerable components for purposes of enhanced monitoring. To be published and presented at IFIP SEC 2011: Future challenges in security and privacy for academia and industry, June 2011 Copyright IFIP SEC 2011: Future challenges in security and privacy for academia and industry, 2011 On Computing Enterprise IT Risk Metrics Sandeep Bhatt and William Horne and Prasad Rao Cloud and Security Lab HP Laboratories 5 Vaughn Drive, Princeton NJ 08540, USA {sandeep.bhatt,william.horne,prasad.rao}@hp.com Abstract. Assessing the vulnerability of large heterogeneous systems is crucial to IT operational decisions such as prioritizing the deployment of security patches and enhanced monitoring. These assessments are based on various criteria, including (i) the NIST National Vulnerability Database which reports tens of thousands of vulnerabilities on individual components, with several thousand added every year, and (ii) the specifics of the enterprise IT infrastructure which includes many components. Assessing the vulnerability of large heterogeneous systems is crucial to IT operational decisions such as prioritizing the deployment of security patches and enhanced monitoring. These assessments are based on various criteria, including (i) the NIST National Vulnerability Database which reports tens of thousands of vulnerabilities on individual components, with several thousand added every year, and (ii) the specifics of the enterprise IT infrastructure which includes many components. Defining and computing appropriate vulnerability metrics to support decision making remains a challenge. Currently, several IT organizations make use of the CVSS metrics that score vulnerabilities on individual components. CVSS does allow for environmental metrics, which are meant to capture the connectivity among the components; unfortunately, within Section 2.3 of [1] there are no guidelines for how these should be defined and, consequently, environmental metrics are rarely defined and used. We present a systematic approach to quantify and automatically compute the risk profile of an enterprise from information about individual vulnerabilities contained in CVSS scores. The metric we propose can be used as the CVSS environmental score. Our metric can be applied to the problem of prioritizing patches, customized to the connectivity of an enterprise. It can also be used to prioritize vulnerable components for purposes of enhanced monitoring.