Server Fault is a question and answer site for system and network administrators. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

First some of the details. We are a SME (80 office users 30 remote) with 2, soon to be 3 WAN connections. Our WAN connections are a T1, Cable (25/5), and soon to be DSL (10,2). We are looking to to provide automatic WAN link failover for inbound services across all of our WAN links.

I feel that some of the WAN link balancing products currently on the market should be able to fit our needs, however I'm looking for any input anyone has on how to best address inbound automatic link failover. I would like to try to stay away from BGP, as our in house expertise is thin in the area.

I did find these three different questions, however most of them didn't really discuss why they felt a particular solution was good or bad.

So I'm looking for your opinions on what the current best approach for inbound automatic link failover. What have you done, or what have you heard is good?

share|improve this question
As written this won't get anything but grief here. However, if you rewrite it along the lines of "what techniques can be used in this situation" rather than product-focused, you'll get a lot better answers. – sysadmin1138 Nov 2 '11 at 20:42
Edited per suggestion. Thanks! – danorth Nov 3 '11 at 0:28

You have to have

  • own IP-space (PI, not PA)
  • own AS (autonomous system)
  • support BGP on your side, configure it correctly, announce your AS to (all) uplinks (agree on this with NOCs), which (announces) your ISPs will forward further to Net
share|improve this answer

I previously used a product called FatPipe Warp (or something named similarly) to do inbound and outbound WAN load balancing. To support inbound load balancing of services, it ran a DNS server across all the WAN interfaces and which you would make as your authoritative. The A records you created would have an IP address on each of the networks and had absurdly short TTLs. The box would then serve the appropriate record depending on link up, weigt, round-robin, etc. For the most part it worked well. My complaints are that the GUI isn't pretty, it was expensive and lacked the ability to track and report protocol/traffic.

You might also want to check out pfsense. It appears to have similar features.

share|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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