02-25-2013 01:20 AM - edited 03-04-2019 07:07 PM
Hi,
The place that I work has a large pipe to the net for normal use, but uses ADSL/Cable lines for guest/customer Internet access - we have about 12 of these.
At the moment we are fining tat some of these links are virtually never used whereas others are saturated at peak times. The hope is to bundle all of these connection into a single virtual link - this should enable us to get rid of a few of them and will same us a tone of money, as they are running at the maximum speed so if we have to provide more capacity we'll be looking at expensive leased lines.
We use mainly 3750's and 6509's in our network and would rather not buy any new hardware if we can avoid it.
As I'm studying SWITCH right now, acronyms like HSRP and GLBP spring to mind but I don't think they are right for this job.
Can someone please point me to the relevant info to do this? Or if you're feeling kind, do me a sample config to have one device loadbalance across two connections to the Internet.
Thanks.
02-25-2013 02:08 AM
Hi,
Here is a good example:
https://supportforums.cisco.com/docs/DOC-8313
Hope it will help.
Best regards,
Abzal
02-25-2013 02:30 AM
Thanks for this, but it's not quite what I was hoping for. I can appreciate the method, but I was hoping for something similar to the way EIGRP would have allowed unequal cost load balancing on an internal network.
I understand there would be limitations in any case - I don't expect one device to be able to have 300Mbs for a single TCP session spread over 12 lines, but I would like the bandwidth shared more inteligently that just splitting it based on protocol or odd/even/etc IP addresses.
02-25-2013 04:51 AM
Ok, then.
Please share your network topology. And what exactly you need.
Hope it will help.
Best regards,
Abzal
02-25-2013 07:58 AM
Thanks. Given the reply after yours I think this might be redundant, but: we currently have DSL/cable connections going into 3750's in a few locations arround the campus - each ISP connection feeds it's own VLAN via a consumer router set as it's own DHCP server, which is then trunked back to a 6509 VSS when is is distributed as needed.
02-25-2013 08:10 AM
Ok, given what was put above I can see I am probably barking up the wrong tree, but it would help my studies, if not the work network, if someone could explain to me why Optimized Edge Routing would be unsuitable for this purpose.
02-26-2013 08:28 AM
Disclaimer
The Author of this posting offers the information contained within this posting without consideration and with the reader's understanding that there's no implied or expressed suitability or fitness for any purpose. Information provided is for informational purposes only and should not be construed as rendering professional advice of any kind. Usage of this posting's information is solely at reader's own risk.
Liability Disclaimer
In no event shall Author be liable for any damages whatsoever (including, without limitation, damages for loss of use, data or profit) arising out of the use or inability to use the posting's information even if Author has been advised of the possibility of such damage.
Posting
Athough OER/PfR is designed for load balancing, from what you describe, you may have some issues to contend with.
First, the feature isn't available on every platform, for example, don't believe it has any support on 3560/3750 switches. It also might require, on some platforms, an upgraded IOS image or feature license.
Second, OER/PfR "programs" your edge routers with "better" paths via another edge router. You have to insure traffic that reaches the non-optimal edge router can be successfully redirected to the optimal edge router; or insure your whole network "knows" what's optimal. This isn't too difficult when the edge routers are adjacent, but could be difficult with your edge routers distributed about your campus. (PfR's PIRO might ease this, but I haven't worked with that feature.)
Third, OER/PfR "programs" for egress. PfR can support ingress balancing too, but then you need to work with your WAN partner(s) as PfR needs to "inform" their devices how you want ingress directed.
Fourth, for many small routers, you'll probably want a more powerful router as the master controller.
With the "right" equipment, you probably could make OER/PfR optimize your egress bandwidth. Ingress optimization, working with low-end ISPs, probably won't be possible.
PS:
BTW, OER/PfR originally supported statics or BGP. I.e. you're not BGP peering with your ISPs isn't necessarily a show stopper. PfR's (later) PIRO supports any routing protocol.
Discover and save your favorite ideas. Come back to expert answers, step-by-step guides, recent topics, and more.
New here? Get started with these tips. How to use Community New member guide