[Forces-protocol] Re: Data encoding -- first part

[Alan DeKok <[e-mail]>]

Jamal Hadi Salim wrote:

> Every thing is a TLV.
> The T is the ID as provided by the XML doc.
> 
> The disadvantage is clearly the abuse of bits. You could play tricks to
> improve this of course. 
> The advantage is that this is a well know interface.

   There are intermediate possibilities.  Rather than putting every 
element of every data structure into a unique TLV, you can have a TLV 
like "LFB-Instance-Data", which contains all of the PACKed information 
that Zsolt is talking about.

   The benefits with this approach are:

- you have one LFB-Instance-Data TLV for all LFB's, rather than multiple 
data-specific TLV's, or multiple LFB-specific LFB's.

- The protocol is more compact, efficient, and easy to parse by things 
that don't understand the internals of LFB configuration.

- you can update the PACKed data and STRUCTs inside of LFB-Instance-Data 
without adding or deleting TLV's.


   I would propose the above scheme, and also a TLV such as 
LFB-Data-Structure, which could hold an encoded representation of the 
STRUCTs and data that the LFB understands, so that the CE can 
"bootstrap" it's knowledge of how to configure the LFB.

   This process could be very powerful.

   Alan DeKok.



_______________________________________________
Forces-protocol mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/forces-protocol