[Sis-dtn] [EXTERNAL] RE: New version of LTP Corrigendum
Felix Flentge
Felix.Flentge at esa.int
Thu Jun 1 14:34:49 UTC 2023
Hi,
I would like to share what I have already sent to smaller audience (sorry for the repetition; it seems that part of the email thread did not go the mailing list).
What is important to me is that
1. Not-LTPv2 does not impose any restrictions on block sizes and does provide the same service interface for reliable and unreliable sessions:
"In my view, for the new protocol, the main use for unreliable 'not-LTP' is not minimized delay:
* Reliable or unreliable not-LTP with rangeReceived.indication can be used to minimize delay (and with small blocks to fit a segment if you like)
* Use cases for unreliable LTP include situations where you don't have a return link available. This can be cases where you have a very re-liable forward link. This is why we have the same service interface for reliable and un-reliable not-LTP."
2. Not-LTPv2 does not need to have access to a contact plan:
"The decision when to consider an unreliable (non-LTP) session closed may also be implementation-specific (but we should require that something is implemented): it may be based on a timer, an external event (end of a communication session) or the arrival of a segment for a new session (un-reliable or maybe even reliable). However, we should avoid to put any constraints regarding the number of parallel sessions or how segments of different sessions are multi-plexed as this depends on various factors (conops, resource constraints, link characteristics, ...)."
I think this could be solved in the same way as the link state procedures in CFDP (notifications of link state via MIB). They way / events how these link states are changed in the MIB is outside the protocol.
Regards,
Felix
From: Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov>
Sent: Wednesday, May 31, 2023 2:55 PM
To: Jeremy.Mayer at dlr.de; sburleig.sb at gmail.com; Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov>; Felix Flentge <Felix.Flentge at esa.int>; sis-dtn at mailman.ccsds.org; keithlscott at gmail.com
Subject: RE: [Sis-dtn] [EXTERNAL] RE: New version of LTP Corrigendum
I am unconvinced either way, to be honest. I agree with Jeremy forcing the LTP engine to know the contact plan is not great, but having another "stale session timeout" seems an equally non-ideal solution. Would all these problems "go away" if we force in the spec that a green block always be transmitted as a single green segment? But then the block size must be matched to the underlying link MTU (or transmission size used by the mission) which is <=65 kB for transfer frames, I believe. Would that be too restrictive if we sent streaming video over LTP green?
-----------------------------------------------------------------------------------------------------------------------
Marc Sanchez Net (332H)
Telecommunications Engineer
Jet Propulsion Laboratory
Cell: (617) 953-7977<mailto:(617)%20953-7977> | Email: marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>
-----------------------------------------------------------------------------------------------------------------------
From: Jeremy.Mayer at dlr.de<mailto:Jeremy.Mayer at dlr.de> <Jeremy.Mayer at dlr.de<mailto:Jeremy.Mayer at dlr.de>>
Sent: Wednesday, May 31, 2023 1:05 AM
To: sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>; Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>; Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>; sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>; keithlscott at gmail.com<mailto:keithlscott at gmail.com>
Subject: RE: [Sis-dtn] [EXTERNAL] RE: New version of LTP Corrigendum
Hi,
Realistically, I agree, the LTP engine should probably be aware of either the contact plan OR the status of the downlink. That being said, I don't think that's a dependency which we want to force, due to the "perpetual downlink" use-case (ISS). There's nothing stopping me from running a single LTP link from now until the end of time, unaware of the status of the underlying link. Of course, there's the caveat that green sessions may be transmitted into the void, never to be seen again.
That being said, the "last green block" problem is insidious: In a nominal downlink (e.g. with minimal out-of-order arrival and loss), we can use the EOB as a signal that the block should be delivered. However, if the EOB is out-of-order, there's a problem: the application shouldn't have to expect that data from the same session may arrive twice. The only way to really deal with this in the face of out-of-order/missing arrivals is to track the completeness of a green session. If the EOB block is delivered while data is still missing, we ignore it and "hold" the session until a user-specified timeout. In LTPv2, we called this the "stale session timeout". If more data is delivered into one of these pending sessions, it can still be enqueued and delivered upon timeout. This allows us to treat out of order sessions and those where the EOB is missed in the same way.
Thanks,
Jeremy
From: SIS-DTN <sis-dtn-bounces at mailman.ccsds.org<mailto:sis-dtn-bounces at mailman.ccsds.org>> On Behalf Of sburleig.sb--- via SIS-DTN
Sent: Tuesday, May 30, 2023 11:55 PM
To: 'Sanchez Net, Marc (US 332H)' <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; 'Torgerson, J. Leigh (US 332C)' <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>; 'Felix Flentge' <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>; keithlscott at gmail.com<mailto:keithlscott at gmail.com>
Subject: Re: [Sis-dtn] [EXTERNAL] RE: New version of LTP Corrigendum
Marc, good question. My thought on this is that LTP needs to be aware of the "contact plan" in order to know when to pause and resume the "red" timers. Given this, the LTP engine should be able to infer that cessation of green block segment reception is due to the termination of the pass. At that point we have a matter of policy. If it's known that green block transmission is always supposed to happen far enough before the end of the pass to enable complete reception, then the end of the pass signifies that any currently incomplete block is not going to be completed and the block's current incomplete contents can be delivered. If not, then maybe the current state of that incomplete final block should be sustained until the start of the next pass enables further reception of that block's segments.
Of course, sticking to small green blocks that are transmitted in single green segments makes the whole issue go away.
Scott
From: Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>
Sent: Tuesday, May 30, 2023 1:18 PM
To: sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>; Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>; 'Felix Flentge' <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>; keithlscott at gmail.com<mailto:keithlscott at gmail.com>
Cc: Gao, Jay L (US 332C) <jay.l.gao at jpl.nasa.gov<mailto:jay.l.gao at jpl.nasa.gov>>; Richard, Nate J (US 332C) <nathaniel.j.richard at jpl.nasa.gov<mailto:nathaniel.j.richard at jpl.nasa.gov>>
Subject: RE: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
All,
A quick update to the draft LTP corrigendum with a few more items to make sure some of the issues being brought up in this email chain don't fall through the cracks.
Also, a minor note on Scott's comment that "the arrival of the first segment of the next "green" block is a simpler and perhaps more accurate (configuration-free) means of determining that it is time to deliver the current block": How does the receiving LTP engine handle the very last LTP block of a pass? Without a timer, would it ever be released to the application? I hate to add new timers, but I do not see how to get around it in this this corner case.
-----------------------------------------------------------------------------------------------------------------------
Marc Sanchez Net (332H)
Telecommunications Engineer
Jet Propulsion Laboratory
Cell: (617) 953-7977<mailto:(617)%20953-7977> | Email: marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>
-----------------------------------------------------------------------------------------------------------------------
From: sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com> <sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>>
Sent: Tuesday, May 30, 2023 8:13 AM
To: Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>; 'Felix Flentge' <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>; Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: Gao, Jay L (US 332C) <jay.l.gao at jpl.nasa.gov<mailto:jay.l.gao at jpl.nasa.gov>>; Richard, Nate J (US 332C) <nathaniel.j.richard at jpl.nasa.gov<mailto:nathaniel.j.richard at jpl.nasa.gov>>
Subject: RE: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
Guys, a thought on this.
In "red" transmission you expect out-of-order segment arrival on a routine basis, because every retransmitted bundle will arrive out of order. There are timers in the protocol to support the operation of the retransmission procedures.
In "green" transmission you cannot have out-of-order segment arrival if you size your blocks in such a way that each block is transmitted in a single segment. I believe users will typically adopt this model, as "green" data will normally be data for which minimized delay is more important than reliability (otherwise they would use "red" transmission).
In "green" transmission where the blocks are large enough to require transmission in multiple segments you have to wait for an entire block to arrive - or until you are confident that any missing segments were actually lost rather than simply forwarded out of order - before delivering the contents of the block. But there's no re-transmission to avoid, because the transmission is "green", right? If you wanted retransmission you would have used "red".
So in that event I would suggest that the arrival of the first segment of the next "green" block is a simpler and perhaps more accurate (configuration-free) means of determining that it is time to deliver the current block - complete or incomplete - and let the application figure out what to do about the missing data.
Scott
From: Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>
Sent: Tuesday, May 30, 2023 7:54 AM
To: Felix Flentge <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>; sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>; Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: Gao, Jay L (US 332C) <jay.l.gao at jpl.nasa.gov<mailto:jay.l.gao at jpl.nasa.gov>>; Richard, Nate J (US 332C) <nathaniel.j.richard at jpl.nasa.gov<mailto:nathaniel.j.richard at jpl.nasa.gov>>
Subject: Re: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
Thanks Felix -
Agreed - A statement warning about how implementations should be wary of out-of-order deliveries would be useful in the corrigendum..
regards,
Leigh
From: Felix Flentge <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>
Date: Monday, May 29, 2023 at 11:47 PM
To: "Torgerson, Jordan L (332M)" <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>, "sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>" <sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>>, "Sanchez Net, Marc (US 332H)" <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>, "'Dr. Keith L Scott via SIS-DTN'" <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: "Gao, Jay L (US 332C)" <jay.l.gao at jpl.nasa.gov<mailto:jay.l.gao at jpl.nasa.gov>>, "Richard, Nate J (US 332C)" <nathaniel.j.richard at jpl.nasa.gov<mailto:nathaniel.j.richard at jpl.nasa.gov>>
Subject: RE: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
Hi Leigh,
Yes, we also have systematic out-of-order in EO or L2 missions as different physical channels may be used (with maybe different rates and different PDU sizes). With 'implementation issue' I don't mean that it is not important but I would like to avoid making it normative as we want to avoid 'changing the standard' which would require interop testing.
By any means, we should have a statement that in case you may have out-of-order delivery it is recommended to implement timer to wait for out-of-order segments to avoid re-transmission (we will add a similar statement to CFDP for the next release).
Regards,
Felix
From: Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>
Sent: Thursday, May 25, 2023 6:22 PM
To: Felix Flentge <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>; sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>; Sanchez Net, Marc (US 332H) <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: Gao, Jay L (US 332C) <jay.l.gao at jpl.nasa.gov<mailto:jay.l.gao at jpl.nasa.gov>>; Richard, Nate J (US 332C) <nathaniel.j.richard at jpl.nasa.gov<mailto:nathaniel.j.richard at jpl.nasa.gov>>
Subject: Re: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
Thanks Felix -
One comment is that in deep space use of DTN, we would expect that out-of-order delivery would be the rule, rather than the exception - if it takes 40 min RTT to recover a missing segment from Mars, waiting for it to ensure in-order delivery to some application would make ops impossible. (One must design one's applications to be "aware" of the operational environment of course..)
We haven't used both LTP red/green at the same time in our testing and real-world ops as far as I know, but I suspect that recommending that red and green be in different sessions would be a good idea, if nothing else but to make troubleshooting easier! (Nate and Jay may have some thoughts based on our lunar ops with KPLO so I cc:d them..)
regards
Leigh
From: Felix Flentge <Felix.Flentge at esa.int<mailto:Felix.Flentge at esa.int>>
Date: Thursday, May 25, 2023 at 4:56 AM
To: "sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>" <sburleig.sb at gmail.com<mailto:sburleig.sb at gmail.com>>, "Sanchez Net, Marc (US 332H)" <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>, "'Dr. Keith L Scott via SIS-DTN'" <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: "Torgerson, Jordan L (332M)" <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>
Subject: [EXTERNAL] RE: [Sis-dtn] New version of LTP Corrigendum
Hi,
I agree with the modifications below.
Some additional points:
* I would propose to also deprecate Service Data Aggregation (it is currently mandatory). Unless, I am overlooking something, it is not an interoperable mechanism as there is no generic way to determine the length of a client data capsule. Also, for BP/LTP the updated BP standard will describe an aggregation mechanism (CBOR length information + bundle IIRC).
* Should we discourage use of mixed sessions (on the other hand LTP green is optional anyway)?
* For the two additional issues, we could add that this is an implementation issue and that implementation may want to introduce these additional timers in case they (routinely) expect out-of-order delivery
Regards,
Felix
From: SIS-DTN <sis-dtn-bounces at mailman.ccsds.org<mailto:sis-dtn-bounces at mailman.ccsds.org>> On Behalf Of sburleig.sb--- via SIS-DTN
Sent: Thursday, May 18, 2023 1:23 AM
To: 'Sanchez Net, Marc (US 332H)' <marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>>; 'Dr. Keith L Scott via SIS-DTN' <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: 'Torgerson, J. Leigh (US 332C)' <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>
Subject: Re: [Sis-dtn] New version of LTP Corrigendum
Marc, FWIW, I agree about deprecating LTP security and I am likewise skeptical that adding more timers is a good idea; that sounds like a way to work around a design element that hasn't been thought through completely.
Scott
From: SIS-DTN <sis-dtn-bounces at mailman.ccsds.org<mailto:sis-dtn-bounces at mailman.ccsds.org>> On Behalf Of Sanchez Net, Marc (US 332H) via SIS-DTN
Sent: Tuesday, May 16, 2023 7:05 PM
To: Dr. Keith L Scott via SIS-DTN <sis-dtn at mailman.ccsds.org<mailto:sis-dtn at mailman.ccsds.org>>
Cc: Torgerson, J. Leigh (US 332C) <jordan.l.torgerson at jpl.nasa.gov<mailto:jordan.l.torgerson at jpl.nasa.gov>>
Subject: [Sis-dtn] New version of LTP Corrigendum
All,
Please find attached a new version of the LTP corrigendum with some modifications including:
* Comparison between LTP and "the new protocol" has been reduced. This in part motivated by the fact that we have demonstrated ~4 Gbps rates with ION's LTP implementation, which is more than sufficient for deep space links (e.g., even in future trunk lines between Earth and Mars, data rates of 4 Gbps are never exceeded).
* I have added two possible additions to the technical corrigendum based on work done by Brian and people at GRC. They are all optional (MAY) statements and I believe can be implemented without additional managed parameters (and timers).
* Brian has commented on two additional issues (see here<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/22__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33vk9LEw4$> and here<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/24__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33uBoiHHY$>), but those seem to require additional timers that would need to be managed, so I am unconvinced it is worth the extra complexity. Brian, please correct me if I am wrong.
Finally, I think the note at the beginning of Section 3.9 of the current CCSDS LTP spec should be modified to explicitly state that LTP security should not be used and, instead, implementers should rely on security mechanisms provided by other parts of the CCSDS protocol stack, be it SDLS or BPSec. Thoughts on this point?
[Image removed by sender.]<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/24__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33uBoiHHY$>
Defer data retransmission with out-of-order report segments * Issue #24 * nasa/HDTN<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/24__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33uBoiHHY$>
When the network is causing out-of-order segment reception it is possible that one or more synchronous reception reports are received either out-of-order or within a short time window, possibly fol...
github.com
[Image removed by sender.]<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/22__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33vk9LEw4$>
Defer synchronous reception report with out-of-order data segments * Issue #22 * nasa/HDTN<https://urldefense.us/v3/__https:/github.com/nasa/HDTN/issues/22__;!!PvBDto6Hs4WbVuu7!J3wEpyPHlAiGcXO0VDc8uO1zb9BOrpYLw3dw6BHPigpmcy0Q4f8gxX6hHLnQOqg_AVxEg_MYrByNRr1urhAp-KViTuP33vk9LEw4$>
When red part data is segmented and delivered to the receiving engine out-of-order, the checkpoint(s) and EORP can be received before the earlier-in-block data segments. If a synchronous report is ...
github.com
Thanks,
-----------------------------------------------------------------------------------------------------------------------
Marc Sanchez Net (332H)
Telecommunications Engineer
Jet Propulsion Laboratory
Cell: (617) 953-7977<mailto:(617)%20953-7977> | Email: marc.sanchez.net at jpl.nasa.gov<mailto:marc.sanchez.net at jpl.nasa.gov>
-----------------------------------------------------------------------------------------------------------------------
This message is intended only for the recipient(s) named above. It may contain proprietary information and/or protected content. Any unauthorised disclosure, use, retention or dissemination is prohibited. If you have received this e-mail in error, please notify the sender immediately. ESA applies appropriate organisational measures to protect personal data, in case of data privacy queries, please contact the ESA Data Protection Officer (dpo at esa.int<mailto:dpo at esa.int>).
This message is intended only for the recipient(s) named above. It may contain proprietary information and/or protected content. Any unauthorised disclosure, use, retention or dissemination is prohibited. If you have received this e-mail in error, please notify the sender immediately. ESA applies appropriate organisational measures to protect personal data, in case of data privacy queries, please contact the ESA Data Protection Officer (dpo at esa.int<mailto:dpo at esa.int>).
This message is intended only for the recipient(s) named above. It may contain proprietary information and/or protected content. Any unauthorised disclosure, use, retention or dissemination is prohibited. If you have received this e-mail in error, please notify the sender immediately. ESA applies appropriate organisational measures to protect personal data, in case of data privacy queries, please contact the ESA Data Protection Officer (dpo at esa.int).
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/sis-dtn/attachments/20230601/ed9ac6b4/attachment-0001.htm>
More information about the SIS-DTN
mailing list