I had that last one once. The IP guy really wanted to join us, but he didn't have the capability in that facility. He wanted us to go to a tier 3\4 market where he had a ton of capacity. His company's transport team had available capacity, but they didn't want to give it to the IP guy. I thought it was a BS answer at the time, but as time has gone on, I've learned that silos in companies really can be that bad. ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ ----- Original Message ----- From: "Julien Goodwin via NANOG" <nanog@lists.nanog.org> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Julien Goodwin" <nanog@studio442.com.au> Sent: Thursday, April 30, 2026 8:49:47 PM Subject: Re: "Peering Router" On 1/5/26 11:27 am, sronan--- via NANOG wrote:
Lots of networks segregate duties, so they may not want a peering connection a router they primarily use for downstream client connections. Translated, they sell services in that market, but don’t want to transit traffic on that same router.
It doesn't even need to be a "want". We reduced what types of peering we took on AS15169 core routers (and later eliminated all of them entirely) not only for reasons that would fall under "want", but we'd also started running into various memory limitations (ACL programming, etc.) on our core boxes due to the breadth of configs, and eliminating them meant we didn't need to replace line cards or worse in boxes that were otherwise trivially handling their traffic needs. This was all well over a decade ago now, but my vague memory is at the time there were only an extremely small number of sessions with non-Google entities (probably less than five) that were actually turned down because of this instead of just shifted to another router in the area (or possibly nearby metro). The other annoying one you can sometimes hear is that a network won't peer with you in a certain facility even though you can see transport kit that's clearly theirs present. Can be many reasons for that too. _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/J52MV5ET...