[Sis-dtn] [EXT] LTPv2 -- LTP encapsulation point

[Dr. Keith L Scott]

Thanks Tomaso; yes, I think I agree that the underlying link characteristics should drive the design choice (or even the real-time adaptive choice, if that’s possible) of the LTP segment size.  The whole idea is that large bundles (sent as LTP blocks) can be split into smaller segments exactly to allow good fits into the link layer and possibly to facilitate FEC.

                                v/r,

                                --keith

From: Tomaso.deCola at dlr.de <Tomaso.deCola at dlr.de>
Date: Wednesday, October 27, 2021 at 8:42 AM
To: Dr. Keith L Scott <kscott at mitre.org>, marc.sanchez.net at jpl.nasa.gov <marc.sanchez.net at jpl.nasa.gov>
Cc: sis-dtn at mailman.ccsds.org <sis-dtn at mailman.ccsds.org>
Subject: [EXT] LTPv2 -- LTP encapsulation point
Hi Keith, Hi Marc,

just looking at the commented version of what will be the red book for LTPv2 I read that there is a question mark as to how efficiently encapsulate LTP segments into Encapsulation Packets, especially in the case of large packets which might be affected by large packet loss rates. It is certainly true that EPP is not capable of doing any slicing, so that if we have a large LTP segment, we’ll still have a large EPP packet. On the other hand TM natively supports (and USLP should do the same) the possibility of splitting a packet into smaller transfer frames, whose size is ruled by the specific channel coding scheme implemented by the underlying C&S sublayer (that typically guarantee reasonably low frame error rates). As such, I’d see the problem the other way round, the transfer frame can have a maximum frame length (that can be a fixed one) and in turn it’s up to the higher layers to decide how much could the specific segment length be to trade off delay, packet loss, and achievable data rate. Moreover I know that one LTP managed parameter is indeed the maximum segment length, so that in my view its selection is mission specific in order to meet some KPIs.
>From this reasoning, I don’t see in fact a big problem in claiming that one LTP segment is mapped onto 1 EPP segment. Or have I misunderstood your concern? The tuning of the segment size is certainly fundamental for the proper mission design, but I don’t think is a matter of a blue book.

Best Regards,

Tomaso


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/sis-dtn/attachments/20211027/49fa6b5c/attachment.htm>