[NANOG] Purpose of Internap's PNET AS22212
Can anybody shed some light on Internap's PNET AS22212? Specifaly how it relates to their PNAP architecture? Is Internap now doing peering? I was under the impression that their entire business model was based around isolated PNAPs and being a backboneless provider. Attempts at getting an explanation from Internap have been fruitless. CT _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
Thurber, PNET (AS22212) is looked upon by InterNAP's PNAPs (its various ASes) as just another transit provider in the mix. So yes, InterNAP technically peers (under AS22212), but there is no guarantee its various PNAPs would choose that path, depending upon a number of factors. (Disclaimer: This is what I've gathered from public information, not from information obtained under non-disclosure.) Regards, Randy -----Original Message----- From: Thurber [mailto:ct@datagram.com] Sent: Monday, April 28, 2008 12:01 PM To: nanog@nanog.org Subject: [NANOG] Purpose of Internap's PNET AS22212 Can anybody shed some light on Internap's PNET AS22212? Specifaly how it relates to their PNAP architecture? Is Internap now doing peering? I was under the impression that their entire business model was based around isolated PNAPs and being a backboneless provider. Attempts at getting an explanation from Internap have been fruitless. CT _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
As an Internap customer, I have been investigating this as well. I'm not sure when PNET got rolled out, but as best I can tell:: a) It is treated as an optional/another transit provider. E.G. When I have investigated changing routing policies to direct traffic to blocks over specific peers I have been given stats on all transit options and PNET was one of them. b) It's a fairly widespread mesh between PNAP's. I believe I have seen what appears to be private-peering between PNET and some large web properties. c) It certainly seemed like there was routing instability inside of PNET last week. d) I have noticed that the "default best" connection from Internap to various locations seems to have changed away from PNET, where it was going before. My guess is that the PNET internal connectivity is primarily MPLS based, as I have not seen TDM speeds across the PNET links, but rather "Internet" speeds. I'm not sure how PNET relates to PNAP-PNAP speeds and associated SLA's from Internap. --D On Mon, Apr 28, 2008 at 9:13 AM, Randy Epstein <repstein@chello.at> wrote:
Thurber,
PNET (AS22212) is looked upon by InterNAP's PNAPs (its various ASes) as just another transit provider in the mix.
So yes, InterNAP technically peers (under AS22212), but there is no guarantee its various PNAPs would choose that path, depending upon a number of factors.
(Disclaimer: This is what I've gathered from public information, not from information obtained under non-disclosure.)
Regards,
Randy
-----Original Message----- From: Thurber [mailto:ct@datagram.com] Sent: Monday, April 28, 2008 12:01 PM To: nanog@nanog.org Subject: [NANOG] Purpose of Internap's PNET AS22212
Can anybody shed some light on Internap's PNET AS22212? Specifaly how it relates to their PNAP architecture? Is Internap now doing peering? I was under the impression that their entire business model was based around isolated PNAPs and being a backboneless provider. Attempts at getting an explanation from Internap have been fruitless.
CT
_______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
_______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
-- -- Darren Bolding -- -- darren@bolding.org -- _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
It's actually very easy to explain. The InterNAP Flow Control (FCP) was designed to route traffic via the cheapest link available provided the link is within an operating range. Over the years the PNAPs have "preferred" different transit providers based on the underlying contract rate - but what rate could possibly be better than free(1)? So as time passes you will more than likely see more and more of AS22212 in your path. -Max 1. Peering is not free, but much less expensive than paying AT&T. On Mon, Apr 28, 2008 at 9:13 AM, Randy Epstein <repstein@chello.at> wrote:
Thurber,
PNET (AS22212) is looked upon by InterNAP's PNAPs (its various ASes) as just another transit provider in the mix.
So yes, InterNAP technically peers (under AS22212), but there is no guarantee its various PNAPs would choose that path, depending upon a number of factors.
(Disclaimer: This is what I've gathered from public information, not from information obtained under non-disclosure.)
Regards,
Randy
-----Original Message----- From: Thurber [mailto:ct@datagram.com] Sent: Monday, April 28, 2008 12:01 PM To: nanog@nanog.org Subject: [NANOG] Purpose of Internap's PNET AS22212
Can anybody shed some light on Internap's PNET AS22212? Specifaly how it relates to their PNAP architecture? Is Internap now doing peering? I was under the impression that their entire business model was based around isolated PNAPs and being a backboneless provider. Attempts at getting an explanation from Internap have been fruitless.
CT
_______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
_______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
_______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
They have a bunch of transit providers they use, PNET is yet another path that they can use. They have enough traffic that goes PNAP <--> PNAP that I presume it would be cheaper to use their own dark fiber or something than to keep paying the other transit providers.
Thurber <ct@datagram.com> 4/28/2008 12:01 PM >>> Can anybody shed some light on Internap's PNET AS22212? Specifaly how it relates to their PNAP architecture? Is Internap now doing peering? I was under the impression that their entire business model was based around isolated PNAPs and being a backboneless provider. Attempts at getting an explanation from Internap have been fruitless.
CT _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog _______________________________________________ NANOG mailing list NANOG@nanog.org http://mailman.nanog.org/mailman/listinfo/nanog
participants (5)
-
Brandon Palmer -
Darren Bolding -
Max Clark -
Randy Epstein -
Thurber