Today, systems rely on assets connected in cyberspace by interdependent and distributed architectures. They form an enterprise of cooperating systems and elements that network in unprecedented ways to operate missions on world-wide scales. The benefits this brings are accompanied by risks that, in this richly connected environment, there are increased opportunities to exploit vulnerabilities within and across their elements. To assure greater resilience to adverse events, it is necessary to identify critical nodes and links that, without which, could compromise the ability of the system to operate as required. This paper provides a new near linear time way to find globally critical nodes and links from local dependencies between elements that comprise a system. C⃝ 2017 Wiley Periodicals, Inc. Syst Eng 20: 297-306, 2017