The MPLS WG Archive[Date Prev][Date Next][Thread Prev][Thread Next] [Date Index][Thread Index][Author Index][Subject Index] Regarding Resource Class Attribute in CRLDP
Ravi To do include/exclude properly you need 2 vectors or you need semantics associated with the one vector (which we don't put in standards as Kireeti reminded me). But apparently you could have a policy to do it. In a single domain you may get by without signaling the vectors since a route computation can be run on a node doing include and exclude and it is unlikely the attributes will change. In the case of multiple domains without route leaking, where you have to do multiple path selections along the path, you need to carry the a vector which would mean that it has to be consistent between domains. Don > -----Original Message----- > From: Ravi Sankar Mantha [mailto:ravi.mantha@wipro.com] > > > Hi , > This is regarding resource class attribute in MPLS; > I am facing problem in understanding the conversion from Explicit > Inclusion, Explicit Exclusion (as specified in RFC2702, > section 5.6.3) > to > the Resource Class Bit Mask ( RsCls field in the Resource Class Tlv in > CRLDP). > Take the case of Explicit Inclusion, supposing the Bit > Mask is 4 of > bits > (for simplicity). ( in reality it is of 32 bits) > Red - bit '0' (LSB) > Blue - bit '1' > Green - bit '2' > Yellow - bit '4' (MSB). > Suppose there is a CRLSP whose resource class affinity is > INCLUDE - > Blue, this can be encoded in the form of bit map as 0010 ( I assume > that > the '0's are considered as Don't care ?), now this CRLSP can traverse > through links whose Resource Class Attributes can be 0110, 0010, 1111 > which include the color- blue) etc. If this is the way, > Resource Class > Bitmap is interpreted by the CRLDP signaling protocol, I wonder how > Explicit > Exclusion can be encoded as bit map, and interpreted by the Signaling > protocol ? > It would be of great help if somebody can through some light on > this., > Thanks in advance, > Ravi > > >
|
|