"Sharing and Protection in a Single-Address-Space Operating System" Review

From: Tarik Nesh-Nash (tarikn_at_microsoft.com)
Date: Wed Jan 14 2004 - 10:53:11 PST

  • Next message: Joanna Muench: "Review of sharing and protection paper"

    The article proposes a new operating system structure that is based on single address space on wide range addresses. This new structure promises to improve reliability, sharing, integration and performance.

    It starts by citing the advantages and disadvantages of multiple-address spaces and then cites why the single space approach should be a better choice; the old system seems too restrictive.

    Then the document describes the design of a prototype "OPAL ". New concepts were first introduced like the "resource groups" and a focus was made on the decoupling between the protection and other parts of the system.

    The next section describes the implementation of OPAL section, the idea of using one kernel for multiple OSs is not new, it was suggested earlier on "The nucleus of a multiprogramming system" paper. I should admit that a lot of the technical details escaped to me in this section.

    The fifth section describes testing the OPAL system in the Boing CAD. Performance results were very positive; the OPAL matches the needs of the Boing application. I will be interested to know the results on diverse type of applications.

    The next section describes some possible issues with the proposed structure. Other issues that I found on a "D. Hagimont , J. Mossière , Cliff W. Hemming, Jr, Hidden capabilities: towards a flexible protection utility for the internet, Proceedings of the seventh workshop on ACM SIGOPS European workshop: Systems support for worldwide applications, September 09-11, 1996, Connemara, Ireland":

      - The thread belongs to only one resource group at a time, this may be inflexible.

      - The secondary storage management by explicit declaration of "interests" is insufficient.

    Overall, this document is well structured and has a lot of useful information. It was hard to read sometimes, this is mainly due to my unfamiliarity to the subject.

     

    PS. I almost missed that Hank is Henry.


  • Next message: Joanna Muench: "Review of sharing and protection paper"

    This archive was generated by hypermail 2.1.6 : Wed Jan 14 2004 - 10:53:29 PST