| Well,
Vitalink maintains that they removed support for remote Translans responding to
the TEST directive on purpose. I have gone a couple of rounds with Vitalink
about this. The story they're giving me now is that "... it was illegal for
a sync line on a LAN Bridge to respond to a LOOP..."
FYI: Correspondence attached, if you're interested...
I have questioned this rather vigorously, and am awaiting response.
From: US1RMC::"[email protected]" "Bill Wood" 24-OCT-1991 19:32:06.08
To: took::mcpherson
CC:
Subj: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
I recevied your message this a.m., since then, I had a couple of people looking
into this question. Here's the answer...
As per my cohort, Brian Pierce, Bridging Product Specialist:
What you are seeing is to be expected, what we did in the old version
of the s/w was illegal. We are now legal by not responding to the
request across a serial link. We will only respond locally.
I hope this answers the question, it may not be the answer you wanted...
hope I helped...
billw
% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us1rmc.ogo.dec.com; id AA10668; Thu, 24 Oct 91 19:30:56 -0400
% Received: by uucp-gw-1.pa.dec.com; id AA13167; Thu, 24 Oct 91 16:18:17 -0700
% Received: from vitalips.NOC.Vitalink.COM by nocsun.NOC.Vitalink.COM with SMTP id AA10653 (5.65a/IDA-1.4.2-V1.4 for [email protected]); Thu, 24 Oct 91 15:47:33 -070
% Received: by vitalips.NOC.Vitalink.COM (5.65+V1.2/V1.3) id AA01239; Thu, 24 Oct 91 15:53:00 -070
% Date: Thu, 24 Oct 91 15:53:00 -0700
% From: [email protected] (Bill Wood)
% Message-Id: <[email protected]>
% To: took::mcpherson
% Subject: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
From: US1RMC::"[email protected]" "Bill Wood" 25-OCT-1991 13:43:16.97
To: took::mcpherson
CC: [email protected], [email protected]
Subj: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
i received your message of :
----- Begin Included Message -----
>
>As per my cohort, Brian Pierce, Bridging Product Specialist:
>
> What you are seeing is to be expected, what we did in the old version
> of the s/w was illegal. We are now legal by not responding to the
> request across a serial link. We will only respond locally.
>
>I hope this answers the question, it may not be the answer you wanted...
>
Well this begs the question:
"*Why* is this illegal?"
If it's an "802 WAN" , then why not implement the 802 protocols (including
TEST) across it, including serial links?
I suppose I can accept the answer if I could get a better explanation of the
rationale for intentionally disabling this feature...
/doug
----- End Included Message -----
I am having Brian take a look with Engineering, he said something about a
document containing this information.
unfortunately, it may not be until next tuesday before i have an answer for
you, i hope this will be able to live until then.
'til next tuesday...
have a good weekend...
billw
% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us1rmc.ogo.dec.com; id AA22253; Fri, 25 Oct 91 13:41:23 -0400
% Received: by uucp-gw-1.pa.dec.com; id AA13900; Fri, 25 Oct 91 10:09:19 -0700
% Received: from vitalips.NOC.Vitalink.COM by nocsun.NOC.Vitalink.COM with SMTP id AA26970 (5.65a/IDA-1.4.2-V1.4 for [email protected]); Fri, 25 Oct 91 09:34:52 -070
% Received: by vitalips.NOC.Vitalink.COM (5.65+V1.2/V1.3) id AA13671; Fri, 25 Oct 91 09:40:19 -070
% Date: Fri, 25 Oct 91 09:40:19 -0700
% From: [email protected] (Bill Wood)
% Message-Id: <[email protected]>
% To: took::mcpherson
% Subject: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
% Cc: [email protected], [email protected]
From: US1RMC::"[email protected]" "Bill Wood" 30-OCT-1991 16:58:54.20
To: took::mcpherson
CC: [email protected]
Subj: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
in your msg from last week...
----- Begin Included Message -----
Well this begs the question:
"*Why* is this illegal?"
If it's an "802 WAN" , then why not implement the 802 protocols (including
TEST) across it, including serial links?
I suppose I can accept the answer if I could get a better explanation of the
rationale for intentionally disabling this feature...
/doug
----- End Included Message -----
well... according to the document "The Ethernet", "A Local Area Network",
"Data Link Layer and Physical Layer Specifications", Version 2.0 November 1982,
Section 8. Ethernet Configuration Testing Protocol. The Ethernet Configuration
Testing Protocol provides a minimum testing capibility of communication between
stations on an Ethernet. It is the only Client Layer protocol specified in
this document and has the only assigned Ethernet type field value in this
document. All Ethernet stations must support the configuration testing
functions.
Our engineering staff interprets this to mean that the test packet will only
apply to the local port, not the remote port.
my own opinion: it seems as though we are trying to, at a minimum, meet the
specification. whether or not we should have deleted this "feature", I can not
judge.
I hope this answers the quetion?
billw
% ====== Internet headers and postmarks (see DECWRL::GATEWAY.DOC) ======
% Received: by us1rmc.ogo.dec.com; id AA29408; Wed, 30 Oct 91 16:57:27 -0500
% Received: by uucp-gw-1.pa.dec.com; id AA08175; Wed, 30 Oct 91 13:02:47 -0800
% Received: from nocsun.NOC.Vitalink.COM by mescal.NOC.Vitalink.COM with SMTP id AA20910 (5.65c/IDA-1.4.4 for <[email protected]>); Wed, 30 Oct 1991 10:21:43 -080
% Received: from vitalips.NOC.Vitalink.COM by nocsun.NOC.Vitalink.COM with SMTP id AA07129 (5.65a/IDA-1.4.2-V1.4 for [email protected]); Wed, 30 Oct 91 10:17:01 -080
% Received: by vitalips.NOC.Vitalink.COM (5.65+V1.2/V1.3) id AA15292; Wed, 30 Oct 91 10:22:46 -080
% Date: Wed, 30 Oct 91 10:22:46 -0800
% From: [email protected] (Bill Wood)
% Message-Id: <[email protected]>
% To: took::mcpherson
% Subject: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
% Cc: [email protected]
From: TOOK::MCPHERSON "Doug McPherson | DTN 226-5145 | LKG2-2/N1 30-Oct-1991 1709" 30-OCT-1991 17:35:00.96
To: US1RMC::"[email protected]"
CC: MCPHERSON
Subj: Re: Has something changed in the MOP implementation on Translans between 6.9 and 6.10 ?
Bill,
Standards are (to some degree) open to interpretation. I can live with a
vendor reading a standards document, then implementing it differently from
another vendor (who has purportedly implemented the _same_ standards...) Such
is life in the world of standards.
I *do* however have some heartburn when a vendor implements their
interpretation of standard, then later (almost capriciously) changes it as part
of an "enhancement", leaving the customer with *less* functionality than
before!
Translan bridges themselves (in that they use leased, relatively low-bandwidth
wide area network connections over serial lines) violate some very BASIC
assumptions that 802.3 networks (and associated LAN protocol stacks) are built
upon:
- high-bandwidth
- extremely low error rates
- limited geographic extent (propagation delays)
That being the case, I find it _extremely_ difficult to believe that Vitalink
would all of a sudden take the time and effort to DISABLE a (very useful, and
_working_) feature in the name of 'minimal adherance to a standard'.
If the purpose of a Translan Bridge is to create an extended (802.3) LAN, then
each Translan, as a legitimate member of that LAN should implement respond to
the "LOOP" (or TEST) directive, disrespective of its geographic location (i.e.
whether or not it's on the local or remote side of a serial line). Prior to
6.10, Translan bridges *did* participate in this manner.
Please consider this a formal request to "re-enable" the "TEST" feature over
serial links. If there is an appropriate channel that I need to use, please
point me to it.
/doug
|