Re: [tcpm] WGLC for UTO (fwd)

[Alfred Hönes <[e-mail]>]

----- Forwarded message from Alfred Hönes -----

> From: Alfred Hönes <[e-mail]>
> To: [e-mail]
> Cc: [e-mail]
> In-Reply-To: <[e-mail]> from Lars Eggert at Nov "13, 2007 10:22:40" am
> Message-Id: <[e-mail]>
> Date: Tue, 13 Nov 2007 10:56:36 +0100 (MEZ)
> Subject: Re: [tcpm] WGLC for UTO

> On Tue, 13 Nov 2007 10:22:40 +0200, Lars Eggert wrote:
>> Hi,
>> 
>> FYI, I'll shortly submit my working copy as -07.
>> 
>> Thanks to the new submission system, it will be a simple thing to
>> push out an -08 revision before the cutoff, if the WG would like to
>> see further changes.
> 
> Indeed, both: it was nearly instantaneous, and I have a few more nits.
> 
> 
> (1)  Section 3
> 
> I'm still not content with the delineation between "USER TIMEOUT"
> and "USER_TIMEOUT" -- the latter being the per-connection
> instantiation (TCB variable) of the former, according to the first
> hanging list entry in Section 3.
> 
> Accordingly, in the last list entry, the second instance of
> "USER_TIMEOUT" should be changed to "USER TIMEOUT".
>                                                          v
>    CHANGEABLE (Boolean)
> |     Flag that controls whether USER_TIMEOUT (TCP's USER_TIMEOUT
>       parameter) may be changed based on an UTO option received from the
>       other end of the connection.  Defaults to true and becomes false
>       when an application explicitly sets USER_TIMEOUT.
> ---                                                      v
>    CHANGEABLE (Boolean)
> |     Flag that controls whether USER_TIMEOUT (TCP's USER TIMEOUT
>       parameter) may be changed based on an UTO option received from the
>       other end of the connection.  Defaults to true and becomes false
>       when an application explicitly sets USER_TIMEOUT.
> 
> The second-to-next paragraph, starting
> 
>    Before opening a connection, ...
> 
> explicitely deals with setting the per-connection variables (in
> the TCB, presumably by a suitable extension to the socket API).
> Hence, it should refer to setting the "USER_TIMEOUT".
> So please change:
>                    vvvvvvvvvvvvvvvv
>                                      [...].  If an application
> |  explicitly sets the USER TIMEOUT, CHANGEABLE MUST become false, to
>    prevent UTO options from the other end to override local application
>    requests.  [...]
> ---                vvvvvvvvvvvv
>                                      [...].  If an application
> |  explicitly sets USER_TIMEOUT, CHANGEABLE MUST become false, to
>    prevent UTO options from the other end to override local application
>    requests.  [...]
> 
> 
> (2)  Section 3.1
> 
> Below the recommended formula and the explanation of its terms,
> the draft now says:
> 
>    The RECOMMENDED formula results in the maximum of the two advertised
>    values to be adopted for the user timeout of the connection on both
> |  ends, provided they are within the upper and lower limits.  [...]
>          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> This might be misunderstood as referring to a *conditional* assignment,
> only to be performed when this precondition is met; but the formula
> describes an *unconditional* assignment, and a two-sided clipping
> operation on the right-hand side.
> That might be better reflected by saying:
> 
>    The RECOMMENDED formula results in the maximum of the two advertised
>    values to be adopted for the user timeout of the connection on both
> |  ends, after clipping to the preset upper and lower limits.  [...]
>          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> Please also consider replacing the newly introduced "preset" by "local"
> if you want to emphasize the fact that these limits are indeed a local
> matter, which means that the values adopted on the two ends of the TCP
> connection might well be different:
> 
>    The RECOMMENDED formula results in the maximum of the two advertised
>    values to be adopted for the user timeout of the connection on both
> |  ends, after clipping to the local upper and lower limits.  [...]
>          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
> 
> I have tried to find a short phrase not expanding the text further;
> I you prefer, chose more verbose wording.
> 
> 
> Kind regards,
>   Alfred.
> 
> -- 
> 
> +------------------------+--------------------------------------------+
> | TR-Sys Alfred Hoenes   |  Alfred Hoenes   Dipl.-Math., Dipl.-Phys.  |
> | Gerlinger Strasse 12   |  Phone: [phone], Fax: -18         |
> | D-71254  Ditzingen     |  E-Mail:  [e-mail]                     |
> +------------------------+--------------------------------------------+

----- End of forwarded message from Alfred Hönes -----



_______________________________________________
tcpm mailing list
[e-mail]
https://www1.ietf.org/mailman/listinfo/tcpm