[Css-csts] A new RID on Unbuffered Data Delivery procedure

[John Pietras]

Yves,

This slipped past me during the formal  Red-1 review, but it is
important enough that it should be addressed and corrected before Red-2
is released.

 

The first sentence of the second paragraph under 4.6.2.2 (Concepts)
states "As each data unit is generated, it is either immediately
transferred or immediately discarded in case of communication link
congestion." This is the only time that discarding a data unit in case
of congestion is mentioned for the Unbuffered Data Delivery (UDD)
procedure. In particular, it is not at all supported by the normative
portion of the procedure specification. In response to the 'data
available' event, the state table (4-18) gives the (simple) action 'Send
data to underlying communication service'. The reference for this simple
action in table 4-12 are to sections 4.6.3.1.2 and 4.6.3.3. Neither of
these sections mentions any discarding of data in case of congestion,
and in fact don't even make mention of submitting the data to the
underlying communication service. w

 

With explicitly defining a discard mechanism, the desired behavior will
not occur under a state of congestion. Instead the reliable underlying
communication service can be expected to queue the subsequent
TransferDataInvocations while it attempts to transfer all of them.

 

I think what is needed is (a) to have the Transferring Data behavior in
4.6.3.3 expanded to include explicit discussion of the discard mechanism
and (b) also make this explicit in the state table. Regarding the state
table, this could perhaps be as simple as changing "'Send data to
underlying communication service'" to something like "IF NOT 'congested'
THEN 'Send data to underlying communication service' ELSE 'discard
data'", where 'congested' is defined as "The underlying communications
service cannot accept the contents of the transfer buffer because of
congestion" (same as for BDD), and the action 'discard data' references
a paragraph in 4.6.3.3 that explicitly defines this behavior.

 

Best regards,

John

 

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ccsds.org/pipermail/css-csts/attachments/20101124/b5540a19/attachment.htm