Topical to today's issues, WiscNet (and likely others) actively utilize BFD on EBGP peering sessions across peering fabric on various IX's where possible. This helps by detecting issues across layer2 paths without having to wait for the BGP hold time to expire. If anyone would like to enable BFD on our direct sessions, please reach out to us at peering@wiscnet.net, or to me directly. We use minimum-interval 500, multiplier 3 at various IX's with little issue. Cheers, -- Chris Wopat Network Engineer, WiscNet wopat@wiscnet.net 608-210-3965
+1 to this. We (AS53597) will happy enable BFD to anyone that supports. Shoot us an email at noc@hoyosconsulting.com. Perhaps we can add a column to the participants page for a Y/N flag there? -- Andrew Hoyos hoyosa@gmail.com
On Oct 4, 2016, at 4:03 PM, Chris Wopat <wopat@wiscnet.net> wrote:
Topical to today's issues, WiscNet (and likely others) actively utilize BFD on EBGP peering sessions across peering fabric on various IX's where possible.
This helps by detecting issues across layer2 paths without having to wait for the BGP hold time to expire.
If anyone would like to enable BFD on our direct sessions, please reach out to us at peering@wiscnet.net, or to me directly.
We use minimum-interval 500, multiplier 3 at various IX's with little issue.
Cheers, -- Chris Wopat Network Engineer, WiscNet wopat@wiscnet.net 608-210-3965
On 04.10.2016 16:09, Andrew Hoyos wrote:
+1 to this. We (AS53597) will happy enable BFD to anyone that supports. Shoot us an email at noc@hoyosconsulting.com.
Perhaps we can add a column to the participants page for a Y/N flag there?
Meee tooo Email peering@supranet.net (AS4150) and we'll be happy to get this added. -- Colin Baker SupraNet Communications, Inc. (608) 572-7634 colinb@supranet.net This message is subject to the SupraNet Email Confidentiality Policy which is located at http://supranet.net/confidentiality
I believe BFD support is pretty new on BIRD. Is anyone aware of other IXs running BFD with Bird? On Tue, Oct 4, 2016 at 4:21 PM, Colin Baker <colinb@supranet.net> wrote:
On 04.10.2016 16:09, Andrew Hoyos wrote:
+1 to this. We (AS53597) will happy enable BFD to anyone that supports. Shoot us an email at noc@hoyosconsulting.com.
Perhaps we can add a column to the participants page for a Y/N flag there?
Meee tooo
Email peering@supranet.net (AS4150) and we'll be happy to get this added.
-- Colin Baker SupraNet Communications, Inc. (608) 572-7634 colinb@supranet.net
This message is subject to the SupraNet Email Confidentiality Policy which is located at http://supranet.net/confidentiality
-- Jay Hanke CTO Neutral Path Communications 3 Civic Center Plaza, Suite 204 Mankato, MN 56001 (507) 327-2398 mobile jayhanke@neutralpath.net www.neutralpath.net
Call it a Madison thing; AS 3128 would also like to BFD peer with interested parties. I like Andrew's "Y/N" column idea. -Michael -----Original Message----- From: MICE Discuss [mailto:MICE-DISCUSS@LISTS.IPHOUSE.NET] On Behalf Of Colin Baker Sent: Tuesday, October 4, 2016 4:21 PM To: MICE-DISCUSS@LISTS.IPHOUSE.NET Subject: Re: [MICE-DISCUSS] BFD support on BGP sessions On 04.10.2016 16:09, Andrew Hoyos wrote:
+1 to this. We (AS53597) will happy enable BFD to anyone that supports. Shoot us an email at noc@hoyosconsulting.com.
Perhaps we can add a column to the participants page for a Y/N flag there?
Meee tooo Email peering@supranet.net (AS4150) and we'll be happy to get this added. -- Colin Baker SupraNet Communications, Inc. (608) 572-7634 colinb@supranet.net This message is subject to the SupraNet Email Confidentiality Policy which is located at http://supranet.net/confidentiality
Is it implied that BFD is always for bi-lateral sessions only? Or do some routers allow configuring BFD separately from BGP in some useful way? (I don't think mine do.) Colin, does Supranet support BFD for both AS4150 and AS5683, or just the former? -- Richard
On 04.10.2016 20:52, Richard Laager wrote:
Is it implied that BFD is always for bi-lateral sessions only? Or do some routers allow configuring BFD separately from BGP in some useful way? (I don't think mine do.)
Would have to be per session to be useful. I'm not aware of any IXPs doing BFD on route servers - it's something worth discussing, and it may be fine with fairly conservative timers, but I suspect members interested in utilizing BFD will probably also want to set up bilateral with those peers.
Colin, does Supranet support BFD for both AS4150 and AS5683, or just the former?
4150. You can actually remove 5683 from the website - it's not currently connected and we likely won't be bringing it back. -- Colin Baker SupraNet Communications, Inc. (608) 572-7634 colinb@supranet.net This message is subject to the SupraNet Email Confidentiality Policy which is located at http://supranet.net/confidentiality
On 10/05/2016 07:17 AM, Colin Baker wrote:
Would have to be per session to be useful. I'm not aware of any IXPs doing BFD on route servers - it's something worth discussing, and it may be fine with fairly conservative timers, but I suspect members interested in utilizing BFD will probably also want to set up bilateral with those peers.
I'm wondering if routers support configuring a BFD neighbor separately from a BGP neighbor and then using BFD for next-hop validation. I'm guessing not. So for those of you looking at using BFD, in addition to the bilateral peering, you'd need to filter the routes coming from the route servers and drop those of your BFD peers, right? Otherwise, when your session to the peer goes down, you'd still be hearing their routes from the route server.
Colin, does Supranet support BFD for both AS4150 and AS5683, or just the former?
4150. You can actually remove 5683 from the website - it's not currently connected and we likely won't be bringing it back.
I've updated the website. The route servers will need to be updated too, and this frees up 206.108.255.28. -- Richard
On 10/04/2016 8:35 PM, Michael Hare wrote:
Call it a Madison thing; AS 3128 would also like to BFD peer with interested parties. I like Andrew's "Y/N" column idea.
-Michael
+1 on all counts.. For any of the folks also at MadIX, I'll do BFD with you there as well. Can we add BFD as a column on its' KB as well? For MICE though, it would be *great* if one of the route servers could be setup with BFD support that way everyone could benefit and withdraw routes faster once it was rolled out to both. It still wouldn't insure that A can talk directly to B, only bilateral peering can solve that. BIRD appears to have a "passive" setting for it too, so it wouldn't try to do BFD unless the other side initiated. -James
-----Original Message----- From: MICE Discuss [mailto:MICE-DISCUSS@LISTS.IPHOUSE.NET] On Behalf Of Colin Baker Sent: Tuesday, October 4, 2016 4:21 PM To: MICE-DISCUSS@LISTS.IPHOUSE.NET Subject: Re: [MICE-DISCUSS] BFD support on BGP sessions
On 04.10.2016 16:09, Andrew Hoyos wrote:
+1 to this. We (AS53597) will happy enable BFD to anyone that supports. Shoot us an email at noc@hoyosconsulting.com.
Perhaps we can add a column to the participants page for a Y/N flag there?
Meee tooo
Email peering@supranet.net (AS4150) and we'll be happy to get this added.
On Fri, Oct 07, 2016 at 04:30:22PM -0500, James Stahr wrote:
For MICE though, it would be *great* if one of the route servers could be setup with BFD support that way everyone could benefit and withdraw routes faster once it was rolled out to both. It still wouldn't insure that A can talk directly to B, only bilateral peering can solve that. BIRD appears to have a "passive" setting for it too, so it wouldn't try to do BFD unless the other side initiated.
So, something like this? protocol bfd { interface "bce1" { min rx interval 20 ms; min tx interval 50 ms; idle tx interval 300 ms; passive; }; neighbor 206.108.255.3; neighbor 206.108.255.4; neighbor 206.108.255.5; neighbor 206.108.255.6; neighbor 206.108.255.7; neighbor 206.108.255.8; .... } And then likewise for the IPv6 daemon? -- Doug McIntyre <merlyn@iphouse.net> ~.~ ipHouse ~.~ Network Engineer/Provisioning/Jack of all Trades
On Oct 7, 2016, at 4:47 PM, Doug McIntyre <merlyn@IPHOUSE.NET> wrote:
protocol bfd { interface "bce1" { min rx interval 20 ms; min tx interval 50 ms; idle tx interval 300 ms; passive; }; neighbor 206.108.255.3; neighbor 206.108.255.4; neighbor 206.108.255.5; neighbor 206.108.255.6; neighbor 206.108.255.7; neighbor 206.108.255.8; .... }
Based on http://bird.network.cz/?get_doc&f=bird-6.html, I’d suggest something a bit less aggressive like 500ms x 3. I know a number of us here are using similar timer settings with good luck. idle tx would seem to not be required since it’s in passive mode (so it won’t send packets to a neighbor until it gets one first) protocol bfd { interface "bce1" { interval 500 ms; multiplier 3 passive; }; -- Andrew Hoyos hoyosa@gmail.com
On Oct 7, 2016, at 5:07 PM, Andrew Hoyos <hoyosa@gmail.com> wrote:
On Oct 7, 2016, at 4:47 PM, Doug McIntyre <merlyn@IPHOUSE.NET> wrote:
protocol bfd { interface "bce1" { min rx interval 20 ms; min tx interval 50 ms; idle tx interval 300 ms; passive; }; neighbor 206.108.255.3; neighbor 206.108.255.4; neighbor 206.108.255.5; neighbor 206.108.255.6; neighbor 206.108.255.7; neighbor 206.108.255.8; .... }
Based on http://bird.network.cz/?get_doc&f=bird-6.html, I’d suggest something a bit less aggressive like 500ms x 3. I know a number of us here are using similar timer settings with good luck. idle tx would seem to not be required since it’s in passive mode (so it won’t send packets to a neighbor until it gets one first)
protocol bfd { interface "bce1" { interval 500 ms; multiplier 3 passive; };
And for the neighbors, you’d want to tie these to the BGP sessions rather than creating direct sessions inside 'protocol bfd'. If you add the ‘bfd’ switch to BGP sessions, it will dynamically create the BFD sessions, and tie BGP state to BFD. -- Andrew Hoyos hoyosa@gmail.com
participants (8)
-
Andrew Hoyos -
Chris Wopat -
Colin Baker -
Doug McIntyre -
James Stahr -
Jason Hanke -
Michael Hare -
Richard Laager