[Sls-slp] RE: [Sis-ipo] Adding a note to Encapsulation Service 4.2.2.4

[Israel, David J. (GSFC-5670)]

William,

This note means that you need to be able to receive encap packets with any of the various header sizes, in the general case.  If a program knows that they would never be receiving data from anybody using the 8 byte header, they would not have to be able to support it.  A program can also choose never to transmit 8 byte headers.

Dave


________________________________
From: sis-ipo-bounces at mailman.ccsds.org [mailto:sis-ipo-bounces at mailman.ccsds.org] On Behalf Of Walsh, William
Sent: Tuesday, April 21, 2009 12:22 AM
To: 'Greenberg, Edward'; 'Kazz, Greg J'; sls-slp at mailman.ccsds.org; sis-ipo at mailman.ccsds.org
Subject: RE: [Sis-ipo] Adding a note to Encapsulation Service 4.2.2.4

Greg, Dave,

Did this discussion refer to the 2, 4 vs 8 byte encap headers,
with 1, 2 or 4 byte length fields?
Note that most programs that I've talked to explicitly exempt their
contractors from supporting the 8 byte encap header.

If this comment regards the choices of idle packets, i.e. 1-byte vs.
variable - I'm not sure that part is a bandwidth issue.

I apologize that I don't clearly remember the context of the discussion today.

Regards,
william



________________________________
From: sis-ipo-bounces at mailman.ccsds.org [mailto:sis-ipo-bounces at mailman.ccsds.org] On Behalf Of Greenberg, Edward
Sent: Monday, April 20, 2009 9:17 PM
To: Kazz, Greg J; sls-slp at mailman.ccsds.org; sis-ipo at mailman.ccsds.org
Subject: Re: [Sis-ipo] Adding a note to Encapsulation Service 4.2.2.4

I agree with Dave's comment except it should be a normative statement not a NOTE!


On 4/20/09 9:12 PM, "Kazz, Greg J" <greg.j.kazz at jpl.nasa.gov> wrote:
All,

During today's SIS-IPO meeting, Dave Israel made what I believe is a very good suggestion, discovered during interoperability testing IP/IPE over Encapsulation over AOS frames.

The problem with the current wording of the Encapsulation Service Specification is that two independent implementations will not interoperate if the receive side of the interface implements Encapsulation packet size differently with respect to the transmit side.

The proposed note would be added to the Encapsulation Service Pink Sheets, which is a topic for discussion in the SLS-SLP WG meeting on Friday.  The section is 4.2.2.4 Length of Length Field

The proposed note is:

NOTE  -  An implementation on the transmitting end may choose to either use a fixed Encapsulation header size or adaptively/dynamically adjust the size of the Encapsulation Header size according to the payload size in order to optimize bandwidth use (minimize header overhead).  Therefore, an implementation must be capable of receiving encapsulation packets implementing either case.

Greg


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.ccsds.org/pipermail/sls-slp/attachments/20090421/4624d5b5/attachment.html>