Review of Opal

From: Steve Arnold (steve.arnold4_at_verizon.net)
Date: Sun Jan 18 2004 - 19:30:30 PST

  • Next message: Prasanna Kumar Jayapal: "Review of Opal paper - Jeffery Chase et al."

    In the Opal paper, the authors describe their single address space operating
    system. In this system, addressing is permanent, that is, they are never
    re-used. The argument is made that 64 bits should be enough to support this.

     
    Their motivation is a "growing an dimporant class of applications that
    consist of highly interacting tools manipulating a shared database." Such
    systems include CAD and massive design tools. Opal simplifies this sharing
    by providing an easy mechanism to do so: shared memory.
     
    In order to accept this "new" system, programmers need to think differently
    about how they program today. Secondary storage is no different than primary
    storage. Almost none of the authors concepts are new, except that they are
    piecing together many concepts for newer 64 bit systems.
     
    Of course, maybe some people have a good reason to be skeptical. One big
    argument for this approach is its ease of virtual addressing and shared
    memory. It seems, however, that you would still have to map everything back
    to the physical hardware. This is just a different way of abstracting it?


  • Next message: Prasanna Kumar Jayapal: "Review of Opal paper - Jeffery Chase et al."

    This archive was generated by hypermail 2.1.6 : Sun Jan 18 2004 - 19:30:06 PST