Well after months of planning, design review with our device vendor and telco, and of course, after contract signature, we found the design everyone said would work will indeed not.
I had one big MetroE between data centers that was much underutilized and carrier outages every couple months which took down half of production. So we thought swapping the one big pipe for two half-size pipes over geographically different paths would be a good move. We had planned to aggregate the two circuits using Cisco Etherchanneling technology. Now once implemented, the carrier realizes that protocol is not supported.
While we would have still pursued this configuration had we known we had to treat the circuits as distinct logical pipes, this is very frustrating since we thought we went through all the efforts with the parties involved to approve the design. Given the nature of the technology, it would have been very difficult to lab out. Lesson learned is that the IT manager (me) needs to be more diligent in validating designs. We don’t necessarily want to doubt our vendors and our own SMEs, but we have a responsibility to gather proof.