RE: Queries regarding LMP.

["Martin Dubuc" <martin.dubuc@edgeflow.com>]

Zhi-Wei,

As soon as there are O-O-O nodes (all optical) inside the network, there
needs to be a procedure for fault isolation, regardless if there are OEO
nodes at the edge. LMP provides such fault isolation mechanism. If the
fault isolation procedure in LMP is not used, the vendor has to come up
with its own proprietary mechanism.

For SONET gears, as you point out, there is no need for this procedure
since this technology has mechanisms to deal with fault isolation.

Martin

-----Original Message-----
From: Zhi-Wei Lin [mailto:zwlin@lucent.com]
Sent: Wednesday, July 25, 2001 12:35 PM
To: Martin Dubuc
Cc: rajesh@tejasnetworks.com; mpls@UU.NET; jplang@calient.net;
kireeti@juniper.net; azinin@cisco.com; ccamp@ops.ietf.org
Subject: Re: Queries regarding LMP.


Hi Martin, comments in-line


Martin Dubuc wrote:
> 
> Q1) a) & b) LMP is useful for auto-discovery of links and monitoring
of
> control and data link health. To determine whether LMP is useful or
not,
> one has to look at the operational aspect of product deployment and
> maintenance, i.e. what tools the vendor wants to implement to
facilitate
> provisioning and maintenance of their customers network. In a small
> network, a vendor may be able to get away with manual configuration of
> links. In an opaque network, it is also easy to pinpoint link
failures.
> However, in a truly optical network, there needs to be mechanisms to
> isolate faults (which LMP provides).

When you say "truly optical network", do you mean a network that has no
O/E/O (no 3R) at all? or is there some 3R at the edge?

As for mechanisms to isolate faults, I think this capability is already
part of networks. For example, SONET has mechanisms such as trace
identifiers, AIS/RDI signals, and tandem connection points for doing
this. OTN (G.709) also has these defined. In addition, OTN has defined 6
flexible layers for TCM and 2 fixed layer for TCM (section and path). 

I think LMP is not necessary for fault isolation, but it could be
another mechanism that may be used (if one so chooses). 

> 
> Usefulness of LMP is not directly tied to use of link bundles. Link
> bundles are mostly used to reduce the flooding of link information
> generated by the routing protocol.
> 
> Q3) Destination address can be multicast or any valid IP address of
> destination node (most likely the NodeID of the destination node).
> 
> -----Original Message-----
> From: rajesh [mailto:rajesh@tejasnetworks.com]
> Sent: Wednesday, July 25, 2001 1:59 AM
> To: mpls@UU.NET; jplang@calient.net; kireeti@juniper.net;
> azinin@cisco.com
> Subject: Queries regarding LMP.
> 
> Hi,
> 
>      Have a few queries regd LMP.
> 
> Q1)  Regd the applicability of LMP...
> 
> draft-many-gmpls-architecture-00.txt states that
> 
> "   For instance, the traditional IP routing model assumes the
>    establishment of a routing adjacency over each link connecting two
>    adjacent nodes. Having such a large number of adjacencies is not
>    scalable at all. Each node needs to maintain each of its
adjacencies
>    one by one, and link state routing information must be flooded in
>    the topology for each link.
> 
>    To solve this issue the concept of bundling was introduced.
>    Moreover, the manual configuration and control of these links, even
>    if they are unnumbered, becomes totally impractical. The Link
>    Management Protocol (LMP) was specified to solve these problems. "
> 
> a) Can one assume that LMP makes sense ONLY in the case of a large
>      number of parallel links and link bundling...
> 
>       and an implementation of GMPLS neednot have LMP if no link
>       bundling is used ??
> 
>       The above maybe the case when using lower end SONET ADM,
>        DXCs...
> 
> b) Under what other conditions would one require LMP (ie other than
>       link bundling) ?
> 
> Q2) Suppose one does NOT need to use LMP but then one needs to use
>          OSPF extensions to support TE
> (draft-kompella-ospf-gmpls-extensions-01.txt) ,
> 
>         then one may need to do a  correlation of Link Ids between 2
> TE nodes.
> 
>        a) Is there a standard recommended way to do this correlation
??
> 
>              E.g OSPF - Link Local Signaling
>                     or OSPF - link scope LSA ?
> 
> Q3) Section 9 of draft-ietf-mpls-lmp-02.txt   says that LMP messages
>          are encoded as IP packets and also suggests protocol number
> 140.
> 
>           What would be the destination IP address of an LMP packet ??
> 
>            A multicast or is it got from an IGP ??
> 
> Q4) In LMP, if a Link Id correlation has been done,  and If the link
ids
> 
>          on one of the ends or both change... How is this change taken
> into
>          account ??
> 
>          Eg for i = 1,2
>                           Port i, Node 1 connected to Port i, Node 2
> 
>           If one interchanges the connection then one has to
recorrelate
> 
>            the Link Id mappings..
> 
> Thanks in advance,
> Rajesh