The MPLS WG Archive

["kishan" <harikishan.desineni@ericsson.com>]

Q1::

 

Section 3.4.5.1.2 LDP RFC 3036 states that, when a label mapping message

is received with a pathvector TLV and loop is detected, the LSR should transmit

 a label release message carrying a Status TLV to the sending LSR to signal

"Loop Detected".

 

 But,the Label release message(3.5.11) does not include a status TLV in it's

optional List of TLVs? So, what is the way to inform downstream that the

label release was a result of loop detected by the upstream??

 

*****

 

Q2::

 

    Check_Received_Attributes (A.2.6) algorithm is in such a way that the

 Path Vector TLV will be processed only when HopCount TLV is present

in the Label Request/Mapping message. ( CRa.1)

 

Does that mean Loop Detection can be used only along with HopCount

and it can't be used independently ( without HopCount )?? I know that

sometimes (for ATM/FR LSRs) HopCount is a MUST in Label Request

messages and HC may or may not be present in Label mapping messages

based on the Merge type of the LSR.

 

The LDP MIB  "draft-ietf-mpls-ldp-mib-07.txt" has the following

loop detection capabilities.

" mplsLdpLsrLoopDetectionCapable OBJECT-TYPE

SYNTAX INTEGER {

none(1),

other(2),

hopCount(3),

pathVector(4),

hopCountAndPathVector(5)

}"

 

which implies that pathVector(4) should be configurable independent

of anyother loop detection mechanism ( like HopCount). This contradicts

with the description of Check_Received_Attributes (Cra.1) Algorithm

in RFC 3036 which states that PathVector TLV will not be processed

if there is no HopCount TLV in the received message.

 

*******

Thanks,

HariKishan