Security modeling is a critical aspect in protecting computing systems. Different techniques, such as attack trees and attack graphs, among others, have been developed to visualize and analyze potential threats and vulnerabilities. However, these techniques have limitations in capturing the complexity and dynamics of current computing systems, often distributed, networked, virtualized and software defined. To address these limitations, this paper introduces a novel security modeling technique, Attack Block Diagrams (ABD), aiming to provide a more flexible and comprehensive approach to security modeling and performance evaluation. ABD combine elements of attack trees, attack graphs, and reliability block diagrams (RBD) into a novel framework for investigating and assessing potential attack paths in complex computing systems. In this paper a formal definition of ABD and some examples taken from the literature are provided to show the advantages of the proposed formalism against other techniques such as attack trees and attack graphs.