FW: [Sis-csi] Text for Transport Layer Section

[Dai Stanton]

I wouldn't call the application of de-interleaving careful design.
Rather the converse is true - if a (link or physical? - this stuff is at
the boundary) layer interleaves and the other end doesn't de-interleave
then the system engineer is clearly a waste of skin. 

To compare this with packet re-ordering is mischievous bordering on just
silly. We're mostly professional engineers trying to get it right not
bored academics trying to score cheap points. 

TCP supports chatty applications and, consistently with those
applications, is intolerant of gross delays and of gross packet
re-ordering. CFDP, on the other hand, is tolerant of long delays and
gross packet re-ordering (dependant on retransmission mode selected -
though most preference seems to be for the deferred option which is the
most out-of-sequence tolerant option).

Dai

-----Original Message-----
From: sis-csi-bounces at mailman.ccsds.org
[mailto:sis-csi-bounces at mailman.ccsds.org] On Behalf Of Lloyd Wood
Sent: 26 April 2006 00:14
To: Scott Burleigh
Cc: sis-csi at mailman.ccsds.org
Subject: Re: [Sis-csi] Text for Transport Layer Section

At Tuesday 2006-04-25 15:43 -0700, Scott Burleigh wrote:
>Lloyd Wood wrote:
>
>>At Tuesday 2006-04-25 14:57 -0700, Scott Burleigh wrote:
>>
>>>CFDP is designed to run over a deep-space radio link, where the 
>>>laws of physics pretty much assure that if bits arrive at all they 
>>>arrive in the order in which they were transmitted.
>>
>>Modern coding theory and interleaving to ensure robustness to 
>>interference (FEC, LDPC, turbo codes, even interleaved concatenated 
>>codes) pretty much ensures that that claimed in-order delivery of 
>>bits does not, in fact, actually happen. Ever. While still obeying 
>>the laws of physics.
>
>We seem to be talking about different things, Lloyd.  I'm pretty 
>sure that the bits transmitted (by which I mean "radiated", i.e., 
>issued from the radio) arrive at the receiver in the order in which 
>they were transmitted.  That's how I dimly recall electromagnetic 
>signal propagation works.

The (likely non-binary) coding symbols that are used to reconstitute 
those bits are only transmitted and radiated after interleaving. So, 
you get your bits after scrambling and descrambling in the modems at 
each end. Not so much laws of physics, but laws of physics and very 
careful design.

I'd like to be clear as possible in this; any understanding of the 
laws of physics (or the conventions of layering, for that matter...) 
will likely vary between list members, and crossing multiple layers 
with different terminologies doesn't help. (What's the link's 
'bandwidth'? Well, depends what layer you're at; a computer scientist 
will have a different answer to a modem guy.)

But it's interesting that, as you say, CFDP was designed for a single 
'one-hop' radio link, effectively treating IP as a link protocol, and 
assuming an ordered link flow. (How that ordered flow is produced or 
ensured can vary.) Analogous to TCP, which suffers in out-of-order 
delivery on its path. The designs have to be robust against 
out-of-order-delivery, but still treat it as a rare edge case that 
doesn't need to be optimised for.

L.

<http://www.ee.surrey.ac.uk/Personal/L.Wood><L.Wood at surrey.ac.uk> 

_______________________________________________
Sis-CSI mailing list
Sis-CSI at mailman.ccsds.org
http://mailman.ccsds.org/cgi-bin/mailman/listinfo/sis-csi