Re: Rmt Digest, Vol 66, Issue 8

Vincent Roca <[email protected]> Wed, 03 Feb 2010 15:00:46 +0100
Newsgroups gmane.ietf.rmt
Message-ID <[email protected]>
Hi,

I agree with Mark. No matter the reasons that led to the current EXT_FTI 
format
(and backward compatibility is anyway a good reason), do not change 
anything.
Then if we think this 32 bits waste per packet containing this EXT_FTI 
is really
an issue, we can as well specify another FEC Encoding ID to fixe it. It 
is IMHO
a wiser solution.

Cheers,

    Vincent


On 02/01/2010 06:23 PM, Watson, Mark wrote:
> This is not an error. It is this way for backwards compatibility with 
> the Experimental RFCs.
>
> The binary format of the FEC Object Transmission Information was 
> originally defined in the FLUTE RFC (for use in the EXT_FTI LCT Header 
> Extension). Section 5.1.2.1 of RFC3926 defined the FEC Encoding 
> Id-specific portion of the FEC OTI for FEC Encoding Ids 0, 128 and 130 
> (The "Compact" FEC Schemes) and defined the Maximum Source Block 
> Length to be 32 bits. At least FEC Encoding Id 128 has a 32-bit 
> Encoding Symbol Id.
>
> Now, we could decide to abandon backwards-compatibility in favour of 
> efficiency, but this would not be an errata. Also I think the Compact 
> No-Code FEC Scheme has likely been implemented fairly widely according 
> to the existing specifications and so we would be in danger of making 
> some existing implementations non-compliant (or, more accurately, of 
> breaking alignment between IETF and other specifications).
>
> ...Mark
>
>
> On 2/1/10 7:28 AM, "Brian Adamson" <[email protected]> wrote:
>
>     The issue that brought this up was a question from a developer who
>     was confused by this.  They were unclear as to whether the FEC
>     Payload ID picture was in error or the FEC OTI picture was in
>     error.  They were thinking they should perhaps use a 32-bit symbol
>     ID.
>
>     The other issue with leaving it as it is with a "wasted 16-bits"
>     is that the OTI is that it is really a wasted 32-bits since the
>     OTI could be specified as 10 bytes total per the following instead
>     of 14 bytes as currently depicted. (This retains the one-half
>     32-bit word that allows ALC and NORM EXT_FTI to be 32-bit aligned
>     (i.e. w/ 2 byte header extension header).  In principle the
>     wastage is _usually_ not a major issue but we have very loosely
>     coordinated multicast  use cases where we include the EX_FTI on
>     every packet bearing content.
>
>           0                   1                   2                   3
>            0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
>           +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>           |                      Transfer Length
>                              |
>           +
>                                   +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>           |                               |     Encoding Symbol Length
>        |
>           +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>           |   Max. Source Block Length    |
>           +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>
>
>     Brian Adamson
>     [email protected]
>
>
>
>
>
>     On Jan 30, 2010, at 7:45 PM, Luby, Michael wrote:
>
>         I see.  However, I'm not sure that this qualifies as an
>         "errata", or is simply a waste of 16 bits in the FEC OTI.  On
>         the other hand, as it is now, the FEC OTI is essentially the
>         same for all the schemes described in this document (the only
>         difference that I saw is that in some there is a "reserved"
>         field whereas in others this reserved field is the FEC
>         Instance ID), so there is the advantage that they are all the
>         same.  Also the overall FEC OTI is not very long to start with
>         and only needs to be received once per object, so the wastage
>         is pretty minimal.  However, I don't have a strong opinion
>         either way so if people feel that this rises to the level of
>         an error that deserves an errata, that is fine with me.
>
>
>          On 1/30/10 4:32 PM, "Adamson@Itd. Mil"
>         <[email protected]> wrote:
>
>
>             According to the text from RFC 5445 Section 3.2.2.2:
>
>              "Maximum-Source-Block-Length: a non-negative integer,
>             less than 2^^32,
>              indicating the maximum number of source symbols in a
>             source block"
>
>              it's in "symbols"  ...
>
>              So the question is whether a 32-bit symbol id was
>             intended or if the
>              block length should be 16-bits.  I would _guess_ the
>             latter was the
>              intention given this is the _compact_ no FEC scheme?
>
>
>              Brian Adamson
>             [email protected]
>
>
>
>
>              On Jan 30, 2010, at 5:28 PM, Luby, Michael wrote:
>
>             > Isn't max source block Length in units of bytes and ESIs
>             iN units of
>             > symbols?
>             >
>             >
>             > ----- Original Message -----
>             > From: [email protected] <[email protected]>
>             > To: [email protected] <[email protected]>
>             > Sent: Sat Jan 30 12:00:02 2010
>             > Subject: Rmt Digest, Vol 66, Issue 8
>             >
>             > If you have received this digest without all the
>             individual message
>             > attachments you will need to update your digest options
>             in your list
>             > subscription.  To do so, go to
>             >
>             > https://www.ietf.org/mailman/listinfo/rmt
>             >
>             > Click the 'Unsubscribe or edit options' button, log in,
>             and set "Get
>             > MIME or Plain Text Digests?" to MIME.  You can set this
>             option
>             > globally for all the list digests you receive at this point.
>             >
>             >
>             >
>             > Send Rmt mailing list submissions to
>             > [email protected]
>             >
>             > To subscribe or unsubscribe via the World Wide Web, visit
>             > https://www.ietf.org/mailman/listinfo/rmt
>             > or, via email, send a message with subject or body 'help' to
>             > [email protected]
>             >
>             > You can reach the person managing the list at
>             > [email protected]
>             >
>             > When replying, please edit your Subject line so it is
>             more specific
>             > than "Re: Contents of Rmt digest..."
>             > _______________________________________________
>             > Rmt mailing list
>             > [email protected]
>             > https://www.ietf.org/mailman/listinfo/rmt
>             >
>
>
>
>
>           _______________________________________________
>         Rmt mailing list
>         [email protected]
>         https://www.ietf.org/mailman/listinfo/rmt
>
>
>
>
> _______________________________________________
> Rmt mailing list
> [email protected]
> https://www.ietf.org/mailman/listinfo/rmt
>

_______________________________________________
Rmt mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rmt