2

We have 2 P2P (8M X 2) connection in between HQ and Factory with Cisco 2951 routers at both ends. We are running these links with OSPF cost set to 90. Both links are active and running simultaneously for backup and increased bandwidth (2 X 8M = 16M). So hosts from LAN side sending traffic to both the links.

But in the case of one link goes down host interruption occurs for a while during other link takes over all the outbound traffic. I'm wondering why not we combine these WAN links together act as a single interface so that users don't feel the interruption.

I have read several article on Cisco web portal on this but can't decide which technology will suit this scenario most as EtherChannel, MFR, BVI etc?

Is there any easy way to combine these interface together having both the links running at the same time.

  • Are you actually exchanging routes via OSPF with your carrier? Also, what is the WAN protocol you use now? – Ron Maupin Feb 17 '16 at 7:03
  • 1
    You say they are P2P links, but are these P2P links from a service provider or something like dark fiber? If your routers are connected to a service provider, you likely won't be able to use any of the options you are looking at unless the service provider will configure those services as well (even if it is "transparent" your router's peer is a service provider device and not your other router). – YLearn Feb 17 '16 at 7:06
  • @YLearn Thanks for the reply. We have these links from two different service provider. – Sayeed Feb 17 '16 at 7:45
  • @Ron Maupin These links are only intercity Data connectivity. and we are exchanging routes via OSPF to all the routers connected. – Sayeed Feb 17 '16 at 7:46
  • Did any answer help you? if so, you should accept the answer so that the question doesn't keep popping up forever, looking for an answer. Alternatively, you could provide and accept your own answer. – Ron Maupin Aug 7 '17 at 19:27
-1

You certainly can load balance at layer 3 with simple static routes. You'll want to configure ip load-sharing per-packet on each interface. Be sure to enable Cisco Express Forwarding, CEF.

  • 1
    I strongly advise against the use of this command unless all things have been properly considered and the OP fully understands the implications of the command. I suggest you do the same Ron before suggesting others use it. For example, you say "Be sure to enable Cisco Express Forwarding, CEF" but the Cisco documentation (cisco.com/c/en/us/support/docs/ip/border-gateway-protocol-bgp/…) says "enable process switching (or disable fast switching)". It also goes on to say "This can crash a low-end router because the CPU must do all the processing." Finally, it also says... – OzNetNerd Aug 17 '16 at 3:03
  • 1
    "there is potential that the packets may arrive out of order at the destination because differential delay may exist within the network." This Cisco document (cisco.com/en/US/products/hw/modules/ps2033/…) also says "This ensures equal utilization of the links, but is a processor intensive task and impacts the overall forwarding performance. This form of per-packet load balancing is not well suited for higher speed interfaces." As I mentioned above, all thigns should be carefully considered before using this command. – OzNetNerd Aug 17 '16 at 3:07
  • 1
    I'm not creating the FUD, I'm simply quoting Cisco's own documentation. My apologies, I did miss that line about CEF allowing it to be done more quickly - though the very next line says "However, it does imply that you have the extra resources to deal with maintaining CEF entries and adjacencies", so it is somethign that should be considered. The original document might be old, but at the top of the page it says "Updated:Jan 08, 2015" so think it's safe to say the information is still relevant. – OzNetNerd Aug 18 '16 at 1:15
  • 1
    While the router might be capable of doing, I'd still recommend staying away from it. As I mentioned above and as covered in this blog post - cisco2960.over-blog.com/2014/04/cef-and-load-sharing.html - "However, using Per-packet load sharing is not generally recommended, because it most commonly results in out-of-order packets, affecting TCP traffic throughput (since TCP will bother to fix the out-of-order) and UDP data loss (since UDP will not bother to fix the out-of-order) and to make things more scary out-of-order packets might be interpreted as an attack by firewalls." – OzNetNerd Aug 18 '16 at 1:20
  • 1
    It's unfortunate you have not learned anything. I'm not asking you to take my word for it, I'm simply asking you listen to Cisco's recommendations on their own equipment. The document you linked to clearly states: "Also, per-packet load balancing can result in out-of-sequence (OOS) packet delivery errors on some routers, which can cause applications such as VoIP to malfunction. Therefore, per-packet load balancing is not recommended." – OzNetNerd Aug 18 '16 at 1:51
1

As @YLearn writes, if you are going through a service provider, you need to have its cooperation, and with two separate providers, your options aren't good at all.

If you have direct connections via something like dark fiber, you have some options. Since you say you are using P2P, you could use MLPPP. This, and the other options will lag if your interface still shows up while the link is down (can happen with things like metro ethernet).

Another option is to leave it as is and enable BFD (assuming your license permits it, or you get the proper license) to detect a failover faster. This has the advantage of failing faster than most of the other options, even if the interface still shows up when the link is down.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.