Proposing improvements to avoid the CRC compromise and the silent specification

In this paper we examine the issues and benefits associated with the use of cyclic redundancy checks (CRC)s that manifest during the later phases of safety-critical software development projects. We look at some less obvious details that will potentially increase return on investment (ROI), avoid common DO-178B certification planning and design difficulties, as well as, speed development and integration. We propose software build process improvements with promise to improve version integrity, executable integrity, version identification, and executable identification. We examine the role of CRCs in software configuration management (SCM) and propose their potential for partial satisfaction of DO-178B objectives, like traceability. We look at time-saving steps for integrating with systems that use one of three standard algorithms. Of particular interest, we provide a table of residues from faulty algorithms and their likely causes and remedies, and provide overlooked tips for asynchronous communication devices to aid in the late development phase. There are analysis and data collection techniques proposed to support initial data transfer error rate claims in the early assessment of system integrity and availability, and to show the satisfaction of established bounds.