• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

2 DSL links for load share/balancing and (outbound) redundancy...how?

goldboyd

Golden Member
A company I do some work for currently has a single SDSL line going into a Cisco 2600 which goes to a Pix which static maps a few of their external ips to internal ips, and also connects to the rest of their internal network. They are adding a second SDSL line through a second ISP for load sharing/balancing and also for the (outbound) redundancy. I realize 2 t1's and running BGP would be the best bet for this, but this would be too expensive/complicated for this organization. My first thought was that this would be a fairly simple task, just setup policy based routing on the 2600 between the links. Then I realized that there was no way for the 2600 to detect when one of the dsl lines is down, because the ethernet interface connected to one of the DSL modems will always show as up/up as long as the modem is on. Am I right in thinking this won't work, or is there any way around this?

My first thought for a solution to this would be to use some kind of linux/unix box instead of the 2600. Although I have no idea if I can, I'm assuming there is some way to do it. My thought would be to have 3 NICs in it, one for each dsl modem, one connected to the Pix. Then have the nix box somehow split the outgoing packets between both dsl links. Is this possible? If so how could I do it, can I do it with just route statements, or maybe using ipf or ipfw? Maybe have everything from 0.0.0.0 to 127.255.255.255 route out one interface and then have everything else route out the other?does this make sense? What one guy over there would prefer is that all their users traffic goes out one dsl link, and all the traffic from their website goes out the other, then when one of the links goes down both share one link?this sounds a lot more complicated...but is this also possible? Then I would run some kind of cronjob every minute or so to check if the both dsl interfaces are up, and if one is down, make the other one the default route. Does this all sound plausible? Any suggestions/tips/hints? will be greatly appreciated.

Thanks in advance.
 
First of all.... I don't think that you would need BGP to do what you are wanting to do. BGP is useful when you have Internet Access from multiple providers. 2 DSL links would be from 1 provider. I'm pretty sure that if you set both Ethernet links up with IP address on them, that it MAY load balance, I think the Cisco router will do load balancing because they are both equal cost. You can do load balancing on a Linux box as well. I think with the 2.4 kernel you can compile in "Equal Cost Multipath" and just set a route for both of them. Since the route would have equal cost I THINK it will do load balancing.... Go search for Load Balancing on a Linux site... I think you will find it.

Edit: I looked in the Kernel ---- When you are compiling it (With "make menuconfig") if you look under the "Networking" section you will see Advanced Router... Under that you will have "equal cost multipath".
 
I assume you are using some type of SDSL WIC in the 2600? So you still have your DSL lines as serial0/0 and serial0/1 on the router, correct? I've only done load balancing with T1's, but the theory is the same. A few commands and you are set:


configure terminal
ip cef
int s0/0
ip load-sharing per-packet
int s0/1
ip load-sharing per-packet
exit
ip route 0.0.0.0 0.0.0.0 s0/0
ip route 0.0.0.0 0.0.0.0 s0/1
end
wr mem


ip cef <---this is a global command. It turns on Cisco Express Forwarding
ip load-sharing per-packet <---- splits the load over multiple interfaces
ip route 0.0.0.0 0.0.0.0 s0/0 <---- default route out both interfaces
ip route 0.0.0.0 0.0.0.0 s0/1 <---- default route out both interfaces

 
Back
Top