[c-nsp] Per packet Load balancing

[Rodney Dunn]

It's better to draw a picture for scenarios like this.

I think I get what you have done.


Internet -- R1 ---- PVCa ----\
            |                 \--CE -- /28
            R2 ---- PVCb -------/

So to loadbalance from the CE to the internet you
put two default routes on the CE pointing at PVCa and PVCb.

Then to get to the /28 from the internet you put a /28
route on R1 to PVCa and a second pointing at R2 so R1 has
equal cost routes (at least from R1's perspective) to get
to the CE. You put the same /28 route on R2 towards PVCb I
assume.

Given that, it still will not guarantee you have equal
load balancing over those links because all it takes is
one large pair of ip addresses to use all the bandwidth.
My post a few days ago went over this in detail.
One flow (src/dst) ip pair could eat up all the bandwidth
on one of the PVC's.

Now if your spread of traffic is pretty good between the
/28 and the Internet you should see *some* load sharing
but it will not be 50/50 most likely.

Your best bet is to run netflow and see who are the top
talkers and to verify the path they are taking on R1
do "sh ip cef exact-route <src> <dst>".

This is an example where you may have to do per-packet
if you are required to get more granular on the traffic
split.

Rodney





     On Mon, Sep 06, 2004 at 11:38:08PM +0530, Amol Sapkal wrote:
> Hi,
> 
> I got a client who connects to me on L3 on 2 seperate routers.
> I have assigned him a pool of /28 (16 IPs).
> 
> The total traffic requirement of the client is 2 Megs. One of the IPs,
> in the above pool hogs 90% of the total available badnwidth.
> 
> Now, the connectivity of the client router (CE) with my router R1 and
> R2 is via Nortel FR switches.
> Initially, the client used to connect to R1 only. But, due to banwidth
> design limitations on the FR switch, I had to connect it on IP with
> the other router R2.
> What I did was,
> 1.Created a PVC 'Y' from the CE which ran upto R2. Made this 1 Megs
> fat. Cut down the original PVC 'X' between CE and R1 to 1Megs.
> 2. Put 2 default routes on the CE to point to these 2 PVCs. This made
> sure that my outgoing traffic (from the client router) is load
> balanced.
> 3.R1, on which the CE initially pumped in 2 Megs, is the one which
> announces the /24 pool containing the client IPs to the Internet.
> 4. On R1, I put 2 static routes to the /28 client pool. One points to
> the original PVC 'X' and the other pointed to R2 (there is a PVC
> running between R1,R2 )
> 
> So basically, what I am trying to achieve is 'incoming traffic load
> balancing' wrt CE.
> This was needed because, the traffic for the /28 pool entered my
> network on R1 and I do not want the entire 2 Megs to pass to CE
> directly. So I put 2 static routes on R1 so that half the traffic is
> diverted via R2 to CE (with an latency overhead)
> 
> But for some reason, the load balancing is not happening.
> 'ip cef' was enabled on R1 before, which I thought, was causing the
> problems (If I am not wrong, flows are mapped as a  combination of Src
> IP/Destn IP to outgoing interface)
> After disabling ip cef, I still see that the incoming traffic on CE is
> not being load balanced.
> 
> Any suggestions?
> 
> 
> 
> -- 
> Warm Regds,
> 
> Amol Sapkal
> 
> --------------------------------------------------------------------
> An eye for an eye makes the whole world blind 
> - Mahatma Gandhi
> --------------------------------------------------------------------
> _______________________________________________
> cisco-nsp mailing list  cisco-nsp at puck.nether.net
> https://puck.nether.net/mailman/listinfo/cisco-nsp
> archive at http://puck.nether.net/pipermail/cisco-nsp/