Hi Pascal
The discussion about aliases is interesting but we should not be distracted to
much by it until we have resolved to initial and core problem, how modules and
data nodes are identified. Unless the concept of registered Module ID (20 bits)
and assigned Data node ID (10 bits) are accepted, the only reliable identifiers
are the module name and the data node path, both are considered too heavy
weight.
In summary:
With the current draft, we can advertise of setup data node aliases only using:
(module name, data node path -> alias ID) which represent ~100
octets.
If Module ID and Data node ID are supported, we can advertise of setup data
node aliases using:
(module ID, data node ID -> alias ID) which represent ~ 5
octets.
[cid:[email protected]]
Michel Veillette
System Architecture Director
Trilliant Inc.
Tel: 450-375-0556 ext. 237
[email protected]<mailto:[email protected]>
www.trilliantinc.com<http://www.trilliantinc.com/>
From: Pascal Thubert (pthubert) [mailto:[email protected]]
Sent: 13 juin 2015 14:31
To: Michel Veillette
Cc: Alexander Pelov; Andy Bierman; [email protected]; [email protected]
Subject: RE: [6tisch] Reserve space for aliases
Yes!
With the intent that the alias may be in fact that of the peer.
For instance if the server is a constrained node, the (fat) client would have
an additional index to that table that is the deviceID or the groupID of the
server.
The server would indicate in the handshake that it is not capable to handle
someone else's aliases. It could indicate its aliases in the first responses
and from then on the client would use the server's aliases.
Cheers,
Pascal
From: Michel Veillette [mailto:[email protected]]
Sent: samedi 13 juin 2015 08:55
To: Pascal Thubert (pthubert)
Cc: Alexander Pelov; Andy Bierman; [email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>
Subject: RE: [6tisch] Reserve space for aliases
Hi Pascal
If I understand correctly your proposal
* the range 1 to 3 is used to alias a data nodes based on a mapping
table containing (module name or ID, YANG ID, alias ID).
If this is the case, a single ID might be sufficient since it will allow
aliasing of up to 1000 data nodes.
* If I understand correctly, the range 4 to 60 is used to alias
complete modules based on a mapping table containing (module name or ID, alias
ID).
Is it correct?
[cid:[email protected]]
Michel Veillette
System Architecture Director
Trilliant Inc.
Tel: 450-375-0556 ext. 237
[email protected]<mailto:[email protected]>
www.trilliantinc.com<http://www.trilliantinc.com/>
_______________________________________________
6tisch mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/6tisch
