[Sis-ams] AMS RID BNSC-18

[Scott Burleigh]

Next up:

    Define Delivery Point Specification as a separate parameter, to
    clarify the relationship between it and the Delivery Specification.

    The "implicitly associated" service mode of the Delivery Point
    Specification should be made explicit, i.e. a field of the Delivery
    Specification. This allows the definition of the mapping of service
    mode to UTS to be made explicit.  This allows for a generic AMS
    implementation to be developed that is configured with supported
    service modes depending upon available UTS's.

    The service mode should consider a third issue of message delivery,
    that is to say "timeliness".  This can be in the form of a deadline,
    a delivery time with associated allowable jitter (before and
    after).  This can then be used for time-slot architectures, such as
    1553 or TTA.

My proposed disposition is this:

    Not accepted.  The present scheme for mapping delivery point to
    service mode by reference to transport service (UTS) is flexible and
    efficient, and it provides all the capability contemplated above.

    Delivery deadline was originally included in the QoS specification
    but was removed in October 2005 after discussion within the Working
    Group (see email from September 30, 2005).  Since it is a
    specialized capability that is of value in real-time applications,
    propose we define an additional standard User Operation that
    provides this service.

Stuart, a couple of points to pursue here.  First, is there some aspect 
of automatically mapping service mode to UTS that you believe the 
current system should support but doesn't?  If so, can you explain a 
little more fully and maybe give an example?  One comment I'd offer here 
is that implicit (rather than explicit) mapping of QoS to UTS delivery 
point is a feature that I believe was lifted directly from the MTS 
requirements.

Second, the timeliness dimension of message delivery (deadline, jitter) 
is indeed an issue, but I think it's a requirement that can only be 
satisfied by the UTS itself; as such, it's exactly the sort of thing I 
would think we'd want to map flow label into.  Would you accept that as 
a resolution to your second point?  On reconsideration I think that's a 
lot better answer than using a standard User Operation.

Scott


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ccsds.org/pipermail/sis-ams/attachments/20071017/9a0dba22/attachment.html