Hi, as a follow-up item to my previous email:
As previously noted, we do not utilize the 'Group-ID' field on the
transmit side. Also, I've confirmed with the developer that
we do not perform the processing for group withdrawals on the receive side either - at this time.
If you have hard requirements for this functionality, let your Juniper account team know.
Thanks,
Bob O'Hara
-----Original Message-----
From: Metz, Eduard
To: 'Giles Heron'; Robert O'Hara
Cc: fraanro; juniper-nsp@puck.nether.net; juniper@groupstudy.com
Sent: 2/28/2002 9:49 AM
Subject: RE: Martini support in Juniper
Even worse, martini draft specifies that VC-ID is non-zero. So putting 0
in
there would not be allowed.
It makes sense that the "virtual-circuit-id" be mapped to VC-ID field in
the
martini-header. And "encapsulation-type" to VC-TYPE, but apparently this
is
not yet fully supported.
cheers,
Eduard
> -----Original Message-----
> From: Giles Heron [mailto:giles@packetexchange.net]
> Sent: Thursday, 28 February, 2002 18:08
> To: Robert O'Hara
> Cc: fraanro; juniper-nsp@puck.nether.net; juniper@groupstudy.com
> Subject: Re: Martini support in Juniper
>
>
> sounds a bit strange to me.
>
> The VC ID field is used to uniquely identify a circuit of a
> given type
> between two PE devices.
>
> If you always set VC ID to zero then you can only ever have
> one circuit
> of one type between two PEs :(
>
> Giles
>
> Robert O'Hara wrote:
>
> > Hello Javier,
> >
> > I exchanged email with one of our key developers
> > and he pointed out that we do not utilize the 'Group ID' field
> > on the transmit side. He was uncertain as to whether or not
> > we look at it the info on the receive side and I am still
> investigating
> > that question.
> >
> > Also, for the 'VC-ID' field, we leave it marked at '0', and
> > this indicates that we do not make use of this field at all.
> >
> > I'll let you know about the 'Group ID' on the receive
> > side as soon as I hear back from a different developer.
> > Stay tuned.
> >
> > Thanks,
> >
> > Bob O'Hara
> >
> > Systems Engineer
> > Juniper Networks
> > Northeast Region
> >
> > -----Original Message-----
> > From: fraanro
> > To: Robert O'Hara
> > Cc: juniper-nsp@puck.nether.net; juniper@groupstudy.com
> > Sent: 2/27/2002 9:49 AM
> > Subject: Re: RE: Martini support in Juniper
> >
> > That means:
> > a) The encapsulation used for Ethernet circuits is compliant to
> > Martini.
> > b) The signalling used is compliant with Martini.
> >
> > Now, Martini does not specify how to use the Group VC and
> VC ID fields.
> > How does exactly Juniper use them? Is the VC ID field from the
> > configuration mapped to the VC ID and the Group VC left to 0, for
> > example?
> >
> >
> > Thanks.
> > Javier.
> >
> >
> > ----- Mensaje Original -----
> > Remitente: "Robert O'Hara" <rohara@juniper.net>
> > Fecha: Miércoles, Febrero 27, 2002 5:39 pm
> > Asunto: RE: Martini support in Juniper
> >
> >
> >>Hello Javier,
> >>
> >>Prior to 5.0R4, when signalling the kompella-l2vpn, the
> >>encapsulation codepoints were not compliant with what is
> specified in
> >>draft-martini. So, Juniper changed those codepoints to be
> compliant.
> >>This doesn't indicate that we are compliant with the encapsulation
> >>formats specified in draft-martini. In fact, we are not compliant.
> >>
> >>So, what has changed were the values for encaps types
> >>which are carried in the signalling part of draft-kompella.
> >>
> >>Whatever encapsulation types in draft-martini which REQUIRED
> >>carrying of control words are not supported, e.g.
> Frame-relay and ATM
> >>VC-MUX. So, we are compliant with the circuit types which DO NOT
> >>REQUIRE a control word, e.g Ethernet.
> >>
> >>In Junos 5.2, we now support martini-signalling specified
> >>in the draft-martini-trans. However, this is only for signalling.
> >>So, Martini support for the various circuit types does not exist
> >>in Junos 5.2.
> >>
> >>Hope this helps.
> >>
> >>Bob O'Hara
> >>
> >>Systems Engineer
> >>Juniper Networks
> >>Northeast Region
> >>
> >>
> >>-----Original Message-----
> >>From: fraanro
> >>To: juniper-nsp@puck.nether.net
> >>Cc: juniper@groupstudy.com; juniper-nsp@puck.nether.net
> >>Sent: 2/27/2002 8:42 AM
> >>Subject: Martini support in Juniper
> >>Importance: High
> >>
> >>Hi all,
> >>
> >> In JUNOS 5.2 is supposed to support Martini. I see from the
> >>documentation that for a circuit configuration you must configure
> >>the
> >>Virtual Circuit ID. Now the questions are:
> >>- Is this configured paramenter mapped directly to the VC ID
> >>field in
> >>the LDP signalling VC FEC Element?
> >>- Does Juniper's implementation of Martini use the Group ID for
> >>anything?
> >>- Does it use the VC Type values as defined in the draft, that is:
> >> 0x0001 Frame Relay DLCI
> >> 0x0002 ATM AAL5 VCC transport
> >> 0x0003 ATM transparent cell transport
> >> 0x0004 Ethernet VLAN
> >> 0x0005 Ethernet
> >> 0x0006 HDLC ( Cisco )
> >> 0x0007 PPP
> >> 0x8008 CEM [8]
> >> 0x0009 ATM VCC cell transport
> >> 0x000A ATM VPC cell transport
> >>
> >> - Does it support all the encapsulation types or are
> >>there
> >>restrictions?
> >>
> >>Thanks in advance.
> >>Javier.
> >>
> >>
> >>
> >>
> >>
> >>
> >>
> >
> >
> >
> >
>
>
> --
> =================================================================
> Giles Heron Principal Network Architect PacketExchange Ltd.
> ph: +44 7880 506185 "if you build it they will yawn"
> =================================================================
>
This archive was generated by hypermail 2b29 : Mon Aug 05 2002 - 10:42:39 EDT