Strategic Defense System Distributed Operating System R&D (Research and Development) Review and Recommendations

摘要

The Strategic Defense System (SDS) imposes a set of requirements on distributed operating systems that is not met by state-of-the-art systems. In this paper, the key requirements are identified as being real-time support, reliability/fault tolerance, and security. The extent to which these requirements are being addressed by current distributed operating system research is discussed. The three distributed operating system projects that are currently receiving SDIO funds-Alpha, Cronus, and Mach-are reviewed. A fourth project, the V distributed system project of Stanford University, is also highlighted, because of its unique potential for meeting certain SDS needs. Recommendations on the directions in which the SDIO should pursue each of these projects are made. The Office of Naval Research (ONR) Real-Time Computing Initiative, which is addressing some issues critical to the development of the SDS, is described. It is recommended that the SDIO seek to coordinate with the ONR in this effort. The appendix to this paper provides detailed descriptions of the Alpha, Cronus, Mach, and V distributed operating system projects, as well as three other projects noted in the body of the paper: Amoeba, Clouds, and the Heterogeneous Computer Systems (HCS) Project.