Monday, November 12, 2018

ISP BGP politics/procedures?

In the case of BGP/peering issues between two AS's, which ISP is typically responsible to resolve the issue? I'm trying to figure out if this is something I need to escalate with my CSR or not. This is probably a stupid question, but i've never had to approach any of our ISPs regarding issues on/with public internet.

backstory: Last week, I had a site go offline. SDWAN showing huge packet loss on both links (separate /29s from one wireless provider...long story). Traceroute appears to be all clear, iperf throughput is above CIR, though we were focusing on the radio>ISP core as the likely issue because they are a chronic problem. While I'm working with SDWAN TAC, I get an email from ISP1 engineer, pointing out packet loss found between the destination AS and the upstream for this flow, Telia. Traffic flow from ISP1 to ISP2 is ISP1>TWT>Telia>ISP2, so this is *in my opinion* something that ISP2 should look into. I create a ticket, and $noc-eng1 isn't familiar with AS's, but runs a quick traceroute, emails it to me and washes his hands. It's worth mentioning this is my DC (colo).

This morning, I get a followup asking if the loss is still occurring. I reply with the tech notes and state that ISP1 prepended their AS to send us out through another peer, resolving the issue, but we still see the loss if we hunt for it, between Telia and ISP2. The response I got is that since it's only visible from the remote end, it's on ISP1 to reach out to Telia. Since the peer relationship is between Telia and ISP2, wouldn't it fall on them to resolve it though?

TLDR; unidirectional packet loss between desintation AS and their peer. Who should work to resolve the loss?



No comments:

Post a Comment