Comments on draft-salsano-aquila-sls-00.txt

From: Carlos Alberto Kamienski (cak@cin.ufpe.br)
Date: Fri Nov 10 2000 - 14:12:16 CET

  • Next message: Jim Quinn: "RE: [tequila/sls] SLS Information Model"

    Hi

    I have some comments on this draft.

    Carlos

    ---------------

    1. As the main difference between AQUILA and TEQUILA approaches is the
    concept of predefined services, and their use is optional, I think it
    could be better to put section 4 in an appendix for implementaion issues.

    2. Section 3.3.1, page 7 - "DS-Byte"
       Better to refer to this field as DSField, since it is a six bit field
       now, according to draft-ietf-diffserv-new-terms-03.txt

    3. Section 3.3.2, page 7 - ... IP address can have "partial" wildcards...
       Is it meant something like address/prefix-length ?

    4. Section 3.3.2, page 7 - EXTENDED FlowID
       In fact, it means an SLS may be applied to multiple flows as a whole.

    5. Section 3.4, page 8,

       Table 1 - Why is EAR not applicable to STB ?

       m: minimum policed unit
       maybe I am missing some concept, but what is the real meaning of it ?

    More specific issues (envolving the first trial. btw, I would like to
    know more details of it)

    6. Why was it choosen to use qualitative parameters for implementing
    quantitative services ?

    7. Section 4.1, Page 11
       Parameter m, default value 40B. Since IP + UDP + RTP headers = 40B,
    does it mean a packet with no data ?

    8. Section 4.3, Page 13
       I do not have experience developing streaming applications, but I
    cannot foresee the behaviour of such an application using a congestion
    control mechanism.

    9. Section 3.2, Page 6
       The first trial will implement p2a scopes and applications with strict
    performance requirements. I would like to know about the network
    provisioning for this case.



    This archive was generated by hypermail 2b29 : Fri Nov 10 2000 - 14:13:23 CET