Thread: RE: [mpls-linux-general] Inheriting EXP bit in tunnel cascade
Status: Beta
Brought to you by:
jleu
|
From: <Jon...@cs...> - 2004-08-16 00:04:28
|
Dear James > Currently the EXP bits are new for each level of the=20 > hierarchy, this is considered the 'best common practice' by=20 > those in the MPLS WG at the IETF. At one time we discussed the=20 > scenario you are talking about, the general concessus was=20 > that it was a bad idea. I can appreciate/imagine the reasoning of this general concessus, but just wondering if this scenario can be done from existing mpls-linux implementation. > But there is no reason why I cannot=20 > create a flag that allows this sort of operation to occur. =20 > Would you like this to be a flag on the ILM, on the NHLFE, or=20 > a global flag? Personally I would like to have it done in any way convenient for you. But the big issue is that we are still using the old version 1.172 because there is some other linux functionality required not available in 2.6 kernel. Please ignore my request if you have found it inappropriate... Many thanks for your comments... Cheers, Jonathan |
|
From: <Jon...@cs...> - 2004-08-17 03:51:38
|
Dear James, > I think you can convert the EXP bits to tcindex(or nfmark) in=20 > the ILM and then in the NHFLE convert from tcindex(or nfmark)=20 > back to EXP _before_ each push. >=20 > Something like this: >=20 > mplsadm2 -A -I gen:16:0 > mplsadm2 -I gen:16:0 -i=20 > exp2tc:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:pop:fwd:0x2 >=20 > where 0x2 is the key for this NHLFE: >=20 > mplsadm2 -A -O 0 > Key: 0x2 > mplsadm2 -O 0x2 -o=20 > tc2exp:0x3f:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:17:tc2exp > :0xf:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:18:set:eth1:ipv4:1.1.1.1 Silly question... Is there any reason why the mask of first tc2exp is 0x3f while the second one is 0xf ? > I think your feature request is a valid one. AS you can see this > EXP->tcindex, tcindex->EXP is a pain. A simple flag on the=20 > NHLFE would make this process much easier. I agree... Perhaps it is nice to have this option in a newer version, regardless whether it is the best practice or not.=20 Cheers, Jonathan |
|
From: James R. L. <jl...@mi...> - 2004-08-17 13:13:37
|
On Tue, Aug 17, 2004 at 01:49:06PM +1000, Jon...@cs... wrote: > Dear James, > > > I think you can convert the EXP bits to tcindex(or nfmark) in > > the ILM and then in the NHFLE convert from tcindex(or nfmark) > > back to EXP _before_ each push. > > > > Something like this: > > > > mplsadm2 -A -I gen:16:0 > > mplsadm2 -I gen:16:0 -i > > exp2tc:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:pop:fwd:0x2 > > > > where 0x2 is the key for this NHLFE: > > > > mplsadm2 -A -O 0 > > Key: 0x2 > > mplsadm2 -O 0x2 -o > > tc2exp:0x3f:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:17:tc2exp > > :0xf:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:18:set:eth1:ipv4:1.1.1.1 > > Silly question... Is there any reason why the mask of first tc2exp is > 0x3f while the second one is 0xf ? Nope. It's a typo. Technically the mask only needs to be 0x7, but 0xf works just fine. > > I think your feature request is a valid one. AS you can see this > > EXP->tcindex, tcindex->EXP is a pain. A simple flag on the > > NHLFE would make this process much easier. > > I agree... Perhaps it is nice to have this option in a newer version, > regardless whether it is the best practice or not. > > Cheers, > Jonathan > > > ------------------------------------------------------- > SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media > 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 > Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. > http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 > _______________________________________________ > mpls-linux-general mailing list > mpl...@li... > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general -- James R. Leu jl...@mi... |
|
From: James R. L. <jl...@mi...> - 2004-08-16 16:03:17
|
On Mon, Aug 16, 2004 at 10:02:40AM +1000, Jon...@cs... wrote: > Dear James > > > Currently the EXP bits are new for each level of the > > hierarchy, this is considered the 'best common practice' by > > those in the MPLS WG at the IETF. At one time we discussed the > > scenario you are talking about, the general concessus was > > that it was a bad idea. > > I can appreciate/imagine the reasoning of this general concessus, but > just wondering if this scenario can be done from existing mpls-linux > implementation. I think you can convert the EXP bits to tcindex(or nfmark) in the ILM and then in the NHFLE convert from tcindex(or nfmark) back to EXP _before_ each push. Something like this: mplsadm2 -A -I gen:16:0 mplsadm2 -I gen:16:0 -i exp2tc:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:pop:fwd:0x2 where 0x2 is the key for this NHLFE: mplsadm2 -A -O 0 Key: 0x2 mplsadm2 -O 0x2 -o tc2exp:0x3f:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:17:tc2exp:0xf:0:0:1:1:2:2:3:3:4:4:5:5:6:6:7:7:push:gen:18:set:eth1:ipv4:1.1.1.1 The result is that in the ILM we store the EXP bits in the tcindex. Then before we push we set the "current EXP bits" based on the tcindex. After every push the "current EXP bits" are cleared. > > But there is no reason why I cannot > > create a flag that allows this sort of operation to occur. > > Would you like this to be a flag on the ILM, on the NHLFE, or > > a global flag? > > Personally I would like to have it done in any way convenient for you. > But the big issue is that we are still using the old version 1.172 > because there is some other linux functionality required not available > in 2.6 kernel. Please ignore my request if you have found it > inappropriate... I think your feature request is a valid one. AS you can see this EXP->tcindex, tcindex->EXP it a pain. A simple flag on the NHLFE would make this process much easier. > Many thanks for your comments... > > Cheers, > Jonathan > > > ------------------------------------------------------- > SF.Net email is sponsored by Shop4tech.com-Lowest price on Blank Media > 100pk Sonic DVD-R 4x for only $29 -100pk Sonic DVD+R for only $33 > Save 50% off Retail on Ink & Toner - Free Shipping and Free Gift. > http://www.shop4tech.com/z/Inkjet_Cartridges/9_108_r285 > _______________________________________________ > mpls-linux-general mailing list > mpl...@li... > https://lists.sourceforge.net/lists/listinfo/mpls-linux-general -- James R. Leu jl...@mi... |