Generating accurate preemptive security device policy tuning recommendations

Abstract

An approach is provided for determining a likelihood of an attack on a first computer system of a first business. Characteristics are determined for target businesses having target computer systems currently or recently under attack by an entity whose Internet Protocol (IP) address was selected from a list of suspicious IP addresses. Percentages associated with the characteristics are determined. Each percentage indicates a percentage of the target businesses whose associated characteristic matches a corresponding characteristic of the first business. A score is incremented by an amount for each of the percentages that exceeds an associated threshold. The score is incremented by twice the amount if the IP address matches an address of a source or destination of traffic through a security device in the first computer system. A recommendation to change a security policy for the first computer system is generated if the score exceeds twice the predetermined amount.