[Sis-ams] clarification needed?

[Ray, Timothy J. (GSFC-583.0)]

David,


I like your idea of sending a 'you-are-dead' to the node whose death has
been imputed.

 

Tim

 

________________________________

From: Edell, David J. [mailto:David.Edell at jhuapl.edu] 
Sent: Wednesday, September 10, 2008 4:53 PM
To: Ray, Timothy J. (GSFC-583.0); sis-ams at mailman.ccsds.org
Subject: RE: [Sis-ams] clarification needed?

 

I agree that this is an area that can use some further refinement.  

 

My interpretation was that any receipt of a valid 'YOU_ARE_DEAD' packet,
regardless of status, would be to process it and cease all AMS activity
pending a reconnect command from the user-application.  It's description
in 3.1.24 and 4.2.7.4.4 seems to support this.  The "I_AM_STOPPING"
message I had viewed as only being valid during the normal messaging
mode.  

 

I think the usage of the "I_AM_STOPPING" message being sent to the node
that is inferred to be inactive should be deprecated in favor of
directly sending a 'YOU_ARE_DEAD' message.  Any receipt of
'YOU_ARE_DEAD' from its registrar should then be interpreted as valid.  

 

On this topic, re-reading section 4.2.7.4.3, it does not seem to
explicitly state that a message should be transmitted to the terminated
node itself, but rather to every "other" node.  Catching up on my old
messages thoug, I see that you already caught this one. 

 

- David

 

________________________________

From: sis-ams-bounces at mailman.ccsds.org
[mailto:sis-ams-bounces at mailman.ccsds.org] On Behalf Of Ray, Timothy J.
(GSFC-583.0)
Sent: Wednesday, September 10, 2008 2:02 PM
To: sis-ams at mailman.ccsds.org
Subject: [Sis-ams] clarification needed?

Dear WG members,

 

>From an AMS node point-of-view, there are two incoming MPDUs that tell
it to reset to an 'unregistered' condition:

        You-are-dead

        i-am-stopping (with my own node-id)

 

I have found it difficult to determine (with confidence) whether those
MPDUs *always* cause the reset, or are context-dependent.  For example,
suppose a node receives a 'you-are-dead' but has not sent a 'reconnect'?
(My interpretation says "ignore it")   Conversely, suppose a node sends
a 'reconnect' and then receives an 'i-am-stopping' (with its own
node-id)?  (My interpretation says "reset to unregistered condition").

 

My suggestion is that we discuss this in Berlin and perhaps tighten up
the spec.

 

Tim

 

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.ccsds.org/pipermail/sis-ams/attachments/20080910/ac105d8f/attachment-0001.htm