cancel
Showing results for 
Search instead for 
Did you mean: 
cancel
870
Views
10
Helpful
7
Replies

old classic routing issue - dual BGP internet router with default route

perkin
Level 1
Level 1

Hello experts,

 

I know this is the topic which discusses with a lot of thread, but I took a look and can't found any perfect solution. the classic topology below
===ISP A AS 101===               ===ISP B AS 102===
               |                                           |
       AS100_rtrA                        AS100_rtrB
               |                                           |
===Company class C IP address AS100===

 

Company direction using BGP with receiving default route and advertise a standard class C, also with IPSLA to change the HSRP for the failover if the main link (RTR A) goes wrong.  we all understand that is just able to influence the outbound but not inbound traffic.

 

But now, I want to provide a better failover by
1) putting a peer link in between to provide additional failover layer. also 
2) base on IP SLA if the link quality goes wrong,(e.g ping lost) then I can swing both i/o bound traffic to rtr_b 

 

===ISP A AS 101===               ===ISP B AS 102===
               |                                           |
       AS100_rtrA--------------AS100_rtrB
               |                                           |
===Company class C IP address AS100===


I read a lot of failover post, but some people doing in a "low-tech" way by using EEM to "type" additional command in CLI (like shutdown BGP instance in the router with problem ISP uplink), also some BGP guru has a lot BGP config in "high-tech", way --but those cant change based on the actual link quality.

Question

Is that any way that I can adjust local preference and AS prepend by based on IP SLA?

 

7 Replies 7

Sergey Lisitsin
VIP Alumni
VIP Alumni

perkin,

 

I think a better solution would be to use floating static route on primary router that would depend on IP SLA probe. If the link gone down, the route would disappear and router would stop advertising it altogether via BGP. Coupled with the same SLA controlling the HSRP failover, you would end up with a very neat and consistent failover scenario.

If using this floating static solution meaning
1) the logic is I need something action when IPSLA *down* ? Floating static route will be remove the primary static route when IP SLA fail
2) this solution can’t trigger as prepend asymmetrical routing

Other than 1 or 2 why I am looking for a BGP layer solution since we may cater to Load entire bgp table next steps

EMM can solve my concern but I like to see if any more formal way solution.

Hello,

 

basically, you can configure the IP SLA to monitor jitter/delay/packet loss and have an EEM script change the local preference based on the status of the IP SLA. That would be the simplest way to accomplish this.

 

That said, you also might want to have a look at PfR (Perforrmance Routing)...

 

https://www.cisco.com/c/en/us/td/docs/ios-xml/ios/pfr/configuration/15-mt/pfr-15-mt-book/pfr-understand.pdf

thanks Georg,

 

thanks for your link for performance routing, I did read and google before my response.  It seems to me that is more required both sides has PFR configured, it seems that may not very fit for my internet edge design.

 

but that is quite useful for my next internal design :-)

 

so in short, I may need to do the "cli" way with EEM :-)

PfR doesn't require the other side to have it. PfR handles egress very well, it's trying to handle ingress that's difficult. For ingress, much depends on what the other side supports interacting with your site, again, they don't need PfR (I recall[?]) to process their egress to your router's PfR directions.

thank you,
if you are an expert in this area, please feed me
I am just watching the video from
https://www.cisco.com/c/en/us/products/ios-nx-os-software/performance-routing-pfr/index.html
in the video it shows both side are PfR enabled.

back to my issue, I did some test and may be I will use
1) Local policy for WAN interface source route to external + IP SLA ICMP-ECHO on WAN interface
2) for 1 which Trigger SYSLOG then EEM to change the local preference or remote network command in BGP instance

For the above solution, since I am not using ICMP-jittter so that is not perfect since I cannot test the WAN quality if 20% packet loss....

"f you are an expert in this area, please feed me"

Sorry, I'm no longer an expert in this area. At one company, I pioneered using OER (and then a bit of PfR - when it came on the scene), but that was almost a decade ago. I will say, I was impressed by this technology. If fact, a bit of a semi joke, our network monitoring engineer complained all the network issues "disappeared" after I got OER up and running on our routers. Of course, the network issues were still there but OER routed around them. (Then we had to figure out a way to still monitor the network, which we did by both monitoring OEM/PfR status messages and sending monitoring traffic to not be redirected by OER.)
Getting Started

Find answers to your questions by entering keywords or phrases in the Search bar above. New here? Use these resources to familiarize yourself with the community: