[Sis-csi] Text for Transport Layer Section

[Scott Burleigh]

Lloyd Wood wrote:

> At Tuesday 2006-04-25 13:48 -0700, Scott Burleigh wrote:
>
>> Lloyd Wood wrote:
>>
>>> " CFDP can run over TCP"
>>>
>>> has anyone ever used CFDP over TCP?
>>
>> I believe a TCP "UT layer" adapter for CFDP has been built and tested 
>> at JPL, but I'm pretty sure it's never been used in any operational 
>> sense.
>
> so this wouldn't meet the "two interoperable independent 
> implementations needed to make a standard" requirement, then?

No, I don't think anybody has yet proposed making a TCP UT layer for 
CFDP into a standard.  It wouldn't be a particularly challenging problem 
if someone did, I guess.

>> Unacknowledged CFDP over TCP isn't quite as outlandish a stack as it 
>> might seem: TCP assures the in-order data arrival that the CFDP 
>> procedures rely on (which on rare occasions produces some surprising 
>> behavior in acknowledged CFDP over UDP),
>
> surprising behaviour? not bugs/design flaws?

No, not bugs, and not really design "flaws", though I'd certainly agree 
that this is a limitation of the design.  A fairly well documented one, 
in fact: Note 5 of section 3.4 of the specification says:

    The assumed minimum underlying quality of service is:

    -        with possible errors in the delivered UT_SDUs;

    -        incomplete, with some UT_SDUs missing;

    -        in sequence; i.e., the delivered UT_SDUs are delivered in
    the order in which they were transmitted.

CFDP is designed to run over a deep-space radio link, where the laws of 
physics pretty much assure that if bits arrive at all they arrive in the 
order in which they were transmitted.  You can run it in other contexts, 
but if you do so it may be helpful to read the specification and know 
what you're doing.  Otherwise -- if, for example, you run over UDP/IP in 
the Internet, where packets can take different paths and therefore 
sometimes arrive out of transmission order -- the behavior of a 
conformant and correctly operating implementation of CFDP may on rare 
occasions surprise you.

Scott
. 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ccsds.org/pipermail/sis-csi/attachments/20060425/93b39b31/attachment.html