hi all, my requirment is bit complex so kindly guide me in this,
in my network there are 3 ospf areas and area 0. every area has 2 hubs and on average 40 spokes. spokes have 2 links, primary and backup. now in some spokes both links are falling on the same hub and in some primary is on hub1 and backup is on hub2. we have used ospf cost to manage the links as primary and backup. now we want to utilize the backup links for some specific application like lotus notes. now our analyst team has told us that bulk of lotus traffic COMES to the spoke rather leaving it, so i m quite sure we need to do more configuration on the hub. i was thinking of object tracking but it would involve too much configuration which will introduce complexity and administrative overhead. Can u guys suggest me any feature or technique that helps me achieve my goal efficiently .
hi, ok here is the diagram, cost that i have mentioned is not the actual cost i am using in my environment, its just to give an idea of how i m using cost for pri n backup links.
5 --> primary
10 --> Backup
Kindly check the diagram and guide me through this.
A few questions.
1. Did you intend to "load-balance" for that specific application only or all network traffic.
2. Are the applications that you want to "load-balance" Client/Server based apps wherein the servers are placed in single location.
1) i dont want to load balance lotus notes, i want to shift its traffic to backup link completely and to primary link only when backup fails,
2) as u guys know that lotus notes work in a client server model so yes.
Any ideas ??
If your looking to "guarantee" performance of something like Lotus Notes, I would suggest you do it via a QoS model on any link vs. trying to route just that traffic across a link with "dedicated" or "available" bandwidth. If your primary and backup links are close to the same bandwidth, cost them the same, if they're not, continue to keep them costed differently. Another advantage of using QoS, even if they are costed differently and the primary links fails forcing all traffic to use a lesser bandwidth backup path, QoS should still guarantee Lotus Notes performance.
Dear joseph, we are using qos in our environment but actually we dont want to guarantee much bandwidth to lotus notes, our critical data is our server communications and others, what i want to do is to utilize the backup link, we cant keep the cost same even if the link speed is same, this has been discussed with the management and they strictly prohibit it ( due to some past experiences in our environment they certainly dont want to do load balancing ), i was thinking at random of is there any possible scalable approach to meet my goal, like i said before i came across MTR, do u think it will gonna help me ?
Thanks for the feedback
"MTR"? If you meant, PBR, it could do the job, but it could be maintenance intensive. On the roadmap of PfR, it might support this, but that's some time off. I'm unaware of any other easy way to accomplish your goal, but perhaps someone else will suggest something.