I'm not going to say no. ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs Cogent? On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG <nanog@lists.nanog.org> wrote: Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions? I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP. I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions... ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ...

I’ve said before and i’ll say it again: Friend’s don’t let friends use Cogent -Mike

On Sep 4, 2026, at 09:58, LionLink NOC via NANOG <nanog@lists.nanog.org> wrote:

That’s absurd. IMHO. If NTT is an option, you should talk to someone on their carrier sales team. They’ll build a strategic partnership with you… and last I heard, don’t charge you anything up to a /28.

...

On Fri, Sep 4, 2026 at 12:37 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

I'm not going to say no.

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs

Cogent?

On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q...

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/D3GWGJFM...

This will only increase as IPv4 becomes more and more stupid. I would approach any provider that did so asking if we can simply transit our IPv4 routing process over v6 addresses. This is viable on virtually every common routing platform today. Owen

On Sep 4, 2026, at 09:37, Mike Hammett via NANOG <nanog@lists.nanog.org> wrote:

I'm not going to say no.

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs

Cogent?

On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG <nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q...

I truly feel like Cogent is the Batman villian of our world, but more Tim Burton than Christopher Nolan. On Fri, Sep 4, 2026 at 12:23 PM Owen DeLong via NANOG <nanog@lists.nanog.org> wrote:

This will only increase as IPv4 becomes more and more stupid.

I would approach any provider that did so asking if we can simply transit our IPv4 routing process over v6 addresses. This is viable on virtually every common routing platform today.

Owen

...

On Sep 4, 2026, at 09:37, Mike Hammett via NANOG <nanog@lists.nanog.org> wrote:

I'm not going to say no.

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs

Cogent?

On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ...

...

_______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/56HHGPMV...

-- Chance Callahan Winter: UTC -4 / Summer: UTC -5 Have a nice day!

On Fri, 4 Sept 2026 at 21:47, Brian Knight via NANOG <nanog@lists.nanog.org> wrote: Hey Brian,

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

This might be of interest: https://www.trex.fi/service/turku.html -- ++ytti

[image: image.png] michael brooks Enterprise Network Architect Adams 12 Five Star Schools michael.brooks@adams12.org :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: "flying is learning how to throw yourself at the ground and miss" On Fri, Sep 4, 2026 at 1:36 PM Kevin Tillery via NANOG < nanog@lists.nanog.org> wrote:

I'll be the contrarian here. Surely whether or not this is sane depends on the total price and surrounding circumstances. If a provider charges say $200 for a service, whether they itemize it as $200 connection, free BGP, free IPs; or $180 connection, $20 BGP, free IPs; or $1 connection, $9 BGP, $190 for IPs to run BGP on; is entirely up to them and makes little difference. _______________________________________________ NANOG mailing list

https://urldefense.com/v3/__https://lists.nanog.org/archives/list/nanog@list...

-- This is a staff email account managed by Adams 12 Five Star Schools.  This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender.

An incentive for customers to BYO IPv4 given it’s a limited resource. And because they can. Do not find it shocking.

There's a bit of a bait-and-switch here. I pay for multiple BGP sessions in the contract, but then when it comes time to provision, you want to go to sales to get the IPs to actually provision those sessions. ----- Mike Hammett Intelligent Computing Solutions Midwest Internet Exchange The Brothers WISP ----- Original Message ----- From: "Kevin Tillery via NANOG" <nanog@lists.nanog.org> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Kevin Tillery" <nanog@immibis.com> Sent: Friday, September 4, 2026 2:35:42 PM Subject: Re: Charging for IPs On 4 September 2026 18:30:37 CEST, Mike Hammett via NANOG <nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ...

I'll be the contrarian here. Surely whether or not this is sane depends on the total price and surrounding circumstances. If a provider charges say $200 for a service, whether they itemize it as $200 connection, free BGP, free IPs; or $180 connection, $20 BGP, free IPs; or $1 connection, $9 BGP, $190 for IPs to run BGP on; is entirely up to them and makes little difference. _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/Z4AYTAM5...

They're banking on you already having hardware there and having a cost to move. Or provisioning costs on a new provider if it's a circuit to you. -----Original Message----- From: Mike Hammett via NANOG <nanog@lists.nanog.org> Sent: Saturday, September 5, 2026 3:29 AM To: North American Network Operators Group <nanog@lists.nanog.org> Cc: Mike Hammett <nanog@ics-il.net> Subject: Re: Charging for IPs Also, what's the likelihood that I (or anyone else) would renew the contract when it expires, or add additional ports if that's how I'm treated? It seems cannibalistic. ----- Mike Hammett Intelligent Computing Solutions Midwest Internet Exchange The Brothers WISP ----- Original Message ----- From: "Mike Hammett via NANOG" <nanog@lists.nanog.org> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:30:37 AM Subject: Charging for IPs Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions? I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP. I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions... ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/UQVKYEJE...

Friend’s don’t let friends use Cogent

This is the only answer. On Fri, Sep 4, 2026 at 1:09 PM Mike Lyon via NANOG <nanog@lists.nanog.org> wrote:

I’ve said before and i’ll say it again:

Friend’s don’t let friends use Cogent

-Mike

...

On Sep 4, 2026, at 09:58, LionLink NOC via NANOG <nanog@lists.nanog.org> wrote:

That’s absurd. IMHO. If NTT is an option, you should talk to someone on their carrier sales team. They’ll build a strategic partnership with you… and last I heard, don’t charge you anything up to a /28.

...

On Fri, Sep 4, 2026 at 12:37 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

I'm not going to say no.

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs

Cogent?

On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ...

...

_______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/D3GWGJFM... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/QS5QM2RT...

Cogent has good products, just wish their sales practices were a bit better! Their approach is to assign a new rep after X amount of time of no account activity in the CRM — this includes any sort of communication. Moral of the story, if you have a rep you like, just respond to their emails every so often, even if it’s with a “I’m good, don’t need anything”, so you don’t get a failed used car salesman like this guy assigned to you.

On Sep 5, 2026, at 10:38, Shawn L via NANOG <nanog@lists.nanog.org> wrote:

 It seems like they're stepping up their sales -- We were just assigned a "new rep" though we've never done business with them. He started sending meeting requests, which we declined. Then he would call and actually yell at people because "we had a meeting and we're all waiting for you to join".

Do they really think that this will help them get customers? Even if I was inclined to buy from them, I would not after that.

Shawn

-----Original Message----- From: "Tom Beecher via NANOG" <nanog@lists.nanog.org> Sent: Saturday, September 5, 2026 9:29am To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "LionLink NOC" <noc@lionlink.net>, "Tom Beecher" <beecher@beecher.cc> Subject: Re: Charging for IPs

...

Friend’s don’t let friends use Cogent

This is the only answer.

...

On Fri, Sep 4, 2026 at 1:09 PM Mike Lyon via NANOG <nanog@lists.nanog.org> wrote:

I’ve said before and i’ll say it again:

Friend’s don’t let friends use Cogent

-Mike

...

On Sep 4, 2026, at 09:58, LionLink NOC via NANOG <nanog@lists.nanog.org> wrote:

That’s absurd. IMHO. If NTT is an option, you should talk to someone on their carrier sales team. They’ll build a strategic partnership with you… and last I heard, don’t charge you anything up to a /28.

...

On Fri, Sep 4, 2026 at 12:37 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

I'm not going to say no.

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs

Cogent?

On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG < nanog@lists.nanog.org> wrote:

Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions?

I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP.

I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions...

----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ...

...

_______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/D3GWGJFM... _______________________________________________ NANOG mailing list

https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/QS5QM2RT...

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/Y4XI3D7G... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/VWN3ZZJE...

Yes. It’s exactly that easy. Since the NH is only used to resolve a port and (possibly) a MAC address, having IPv6 NLRI on an IPv4 route is no problem. Of course that only applies to routes where the NLRI is the adjacent advertising router. Owen

On Sep 4, 2026, at 11:38, Brian Knight <ml@knight-networks.com> wrote:

Never mind the knowledge gaps in supporting and managing this, systems updates to handle this, educating customers on how to deploy, etc.

I have never heard of something like this being deployed on IP circuits.

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

-Brian

...

On 2026-09-04 12:23, Owen DeLong via NANOG wrote: This will only increase as IPv4 becomes more and more stupid. I would approach any provider that did so asking if we can simply transit our IPv4 routing process over v6 addresses. This is viable on virtually every common routing platform today. Owen

...

...

On Sep 4, 2026, at 09:37, Mike Hammett via NANOG <nanog@lists.nanog.org> wrote: I'm not going to say no. ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/ From: "Josh Luthman" <josh@imaginenetworksllc.com> To: "North American Network Operators Group" <nanog@lists.nanog.org> Cc: "Mike Hammett" <nanog@ics-il.net> Sent: Friday, September 4, 2026 11:32 AM Subject: Re: Charging for IPs Cogent? On Fri, Sep 4, 2026, 12:30 PM Mike Hammett via NANOG <nanog@lists.nanog.org> wrote: Has anyone else experienced an upstream provider that wants to charge you for the IPs you need to establish BGP sessions? I understand charging for large blocks, but I literally just got a quote for two /31s. I have an existing contract elsewhere with them, and they gave me 3x /29s gratis (for three different ASNs, sets of BGP sessions, ec.). Now they're charging for each IP. I thought charging for BGP sessions was ridiculous, but now for individual IPs to use those BGP sessions... ----- Mike Hammett FD-IX https://fd-ix.com/ DNA Communications https://www.dnacom.com/

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/64BRVJOZ... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/JS2RKN3Q...

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/56HHGPMV...

On Sun, 6 Sept 2026 at 07:04, Owen DeLong via NANOG <nanog@lists.nanog.org> wrote:

Of course that only applies to routes where the NLRI is the adjacent advertising router.

I don't think RFC makes any such restriction, just like IPv4 NH on IPv4 prefix, eBGP would allow it to be any on-link, and iBGP wouldn't care at all. But you should be resetting next-hop in IXP already to the peer address, so that traffic isn't driven to other networks. -- ++ytti

Since the context here is “charging for a /31 for bgp”, I presumed we were talking point to point private peering. Owen

On Sep 5, 2026, at 23:08, Saku Ytti <saku@ytti.fi> wrote:

On Sun, 6 Sept 2026 at 07:04, Owen DeLong via NANOG <nanog@lists.nanog.org> wrote:

...

Of course that only applies to routes where the NLRI is the adjacent advertising router.

I don't think RFC makes any such restriction, just like IPv4 NH on IPv4 prefix, eBGP would allow it to be any on-link, and iBGP wouldn't care at all.

But you should be resetting next-hop in IXP already to the peer address, so that traffic isn't driven to other networks. -- ++ytti

On Sun, 6 Sept 2026 at 20:17, Owen DeLong <owen@delong.com> wrote:

Since the context here is “charging for a /31 for bgp”, I presumed we were talking point to point private peering.

Seems tautology to stress that ebgp with no options on ipv4 (31) or ipv6 (127) for next hop, has no options for next hop. But I guess repetition is the mother of learning In spirit of repetition, this IPv4 NLRI over IPv6 session makes no behavioural changes to class of next-hops allowable. -- ++ytti

This has been raised on the thread already a few times, by different wording, for example Kevin raised the point in a bit different wording. Everyone charges for everything already, because everyone wants cost+, how this is structured is where marketing and psychology lives. On Mon, 7 Sept 2026 at 09:57, Vasilenko Eduard <vasilenko.eduard@huawei.com> wrote:

Reminder, cloud providers charge for every single IPv4 address. Why network provides should give it for free? It is not network providers have started to charge for IPv4. Ed/

...

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Friday, September 4, 2026 21:55 To: North American Network Operators Group <nanog@lists.nanog.org> Cc: Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

On Fri, 4 Sept 2026 at 21:47, Brian Knight via NANOG <nanog@lists.nanog.org> wrote:

Hey Brian,

...

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

This might be of interest: https://www.trex.fi/service/turku.html

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/DMDSA VSOHF5GZEHMAU2ATLGLHCQAK46A/

-- ++ytti

On Mon, 7 Sept 2026 at 10:28, Gary Sparkes <gary@kisaracorporation.com> wrote:

The interface addresses for an IPv4 BGP session should be included in the BGP session cost.

Someone might agree, someone might disagree. Structuring can be used to drive decision making and policy. One valid argument is, driving market to RFC8950 is beneficial, because edge-interfaces are much harder to protect than infrastructure, and having less edge addresses reduces attack surface. But your argument is equally valid, what ever it may be. And it would be fair to say, based on nanog-ml reaction, that it is marketing mistake not to bundle, market reacts negatively to it. -- ++ytti

For direct peering, RFC1918 is valid and usable, so that could be an option providers give - I have a few sessions where I'm using RFC1918 addresses that the provider allocated to me for that first hop session. -----Original Message----- From: Saku Ytti <saku@ytti.fi> Sent: Monday, September 7, 2026 3:33 AM To: Gary Sparkes <gary@kisaracorporation.com> Cc: North American Network Operators Group <nanog@lists.nanog.org>; Vasilenko Eduard <vasilenko.eduard@huawei.com> Subject: Re: Charging for IPs On Mon, 7 Sept 2026 at 10:28, Gary Sparkes <gary@kisaracorporation.com> wrote:

The interface addresses for an IPv4 BGP session should be included in the BGP session cost.

Someone might agree, someone might disagree. Structuring can be used to drive decision making and policy. One valid argument is, driving market to RFC8950 is beneficial, because edge-interfaces are much harder to protect than infrastructure, and having less edge addresses reduces attack surface. But your argument is equally valid, what ever it may be. And it would be fair to say, based on nanog-ml reaction, that it is marketing mistake not to bundle, market reacts negatively to it. -- ++ytti

I was not aware that much (if any?) IPv4 code cared about PMTUD - I thought that only gained provenance in IPv6 Outside of setting DF I guess? But I've had no issues with it, but I only accept/transit 1500 MTU so perhaps I don't see an issue where one might exist -----Original Message----- From: Saku Ytti <saku@ytti.fi> Sent: Monday, September 7, 2026 3:42 AM To: Gary Sparkes <gary@kisaracorporation.com> Cc: North American Network Operators Group <nanog@lists.nanog.org>; Vasilenko Eduard <vasilenko.eduard@huawei.com> Subject: Re: Charging for IPs On Mon, 7 Sept 2026 at 10:38, Gary Sparkes <gary@kisaracorporation.com> wrote:

For direct peering, RFC1918 is valid and usable, so that could be an option providers give - I have a few sessions where I'm using RFC1918 addresses that the provider allocated to me for that first hop session.

RFC1918 breaks PTMUD pretty reliably. You can in many platforms decide to originate ICMP messages from e.g. loopback instead of the RFC1918 interface. But in that case, why bother with the RFC1918, far-side may route it internally and create attack vector to protect against that you have to consider all kinds of attacks towards that RFC1918. With RFC8950 you don't have to, you only have to worry about the IPv6. If you don't have IPv6 and you can originate ICMP from loop, maybe IPv4 LL address is less risky than RFC1918? -- ++ytti

On Mon, 7 Sept 2026 at 10:55, Saku Ytti <saku@ytti.fi> wrote:

CDNs don't use MSS 1460, because it works so poorly. So we are leaving performance on the table, due to PMTUD being too broken.

I seem to have outdated information, most or all(?) have returned to 1460 cap. Largely due to PPPoE density getting lower, combined with PLPMTUD. https://blog.cloudflare.com/client-dynamic-path-mtu-discovery/ -- ++ytti

Except that almost no one charges $405 for transit. They charge $400, whether it comes with BGP and IPs or not. ----- Mike Hammett Intelligent Computing Solutions Midwest Internet Exchange The Brothers WISP ----- Original Message ----- From: "Christopher Hawker via NANOG" <nanog@lists.nanog.org> To: nanog@lists.nanog.org Cc: "Gary Sparkes" <gary@kisaracorporation.com>, nanog@lists.nanog.org, "Christopher Hawker" <chris@thesysadmin.au> Sent: Monday, September 7, 2026 2:36:13 AM Subject: Re: Charging for IPs I can either pay $400/month for a 1G transit service and $5/month for the /31 used by the BGP session, or I can pay $405/month for the transit service with a /31 free. Either way, I’m still paying a total of $405/month. Only benefit to the former is if my upstream will allow me to use my address space for the PtP link, saving $5/month. Regards, Christopher Hawker Sent from my iPhone

On 7 Sep 2026, at 12:59 pm, Gary Sparkes via NANOG <nanog@lists.nanog.org> wrote:

The interface addresses for an IPv4 BGP session should be included in the BGP session cost.

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Monday, September 7, 2026 3:03 AM To: Vasilenko Eduard <vasilenko.eduard@huawei.com> Cc: North American Network Operators Group <nanog@lists.nanog.org>; Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

This has been raised on the thread already a few times, by different wording, for example Kevin raised the point in a bit different wording.

Everyone charges for everything already, because everyone wants cost+, how this is structured is where marketing and psychology lives.

...

On Mon, 7 Sept 2026 at 09:57, Vasilenko Eduard <vasilenko.eduard@huawei.com> wrote:

Reminder, cloud providers charge for every single IPv4 address. Why network provides should give it for free? It is not network providers have started to charge for IPv4. Ed/

...

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Friday, September 4, 2026 21:55 To: North American Network Operators Group <nanog@lists.nanog.org> Cc: Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

On Fri, 4 Sept 2026 at 21:47, Brian Knight via NANOG <nanog@lists.nanog.org> wrote:

Hey Brian,

...

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

This might be of interest: https://www.trex.fi/service/turku.html

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/ DMDSA VSOHF5GZEHMAU2ATLGLHCQAK46A/

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/G7Z4ACHN... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/B2XDYF4F...

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/DI7TUOYO...

On Mon, 7 Sept 2026 at 15:05, Tore Anderson via NANOG <nanog@lists.nanog.org> wrote:

Any ICMPv4/v6 errors can be sourced from the loopback interface, as you say.

We loose interface information in traceroutes if use loopbacks for ICMP errors. I'd immagine people have different opinions about whether this is a problem or not. Personally, since I'm often in the position to troubleshoot problems in other networks, I would rather have this information at hand. With a loopback return IP I cannot even see different ECMP paths if the nodes are the same. I understand that IPv4 has a cost. But using a loopbacks for ICMP errors arguably also does; it's just someone else that is picking up the tab, when troubleshooting issues in the network that is saving a buck. Networks have NOT redistributed public point-to-point prefixes in their IGP for decades to avoid attacks against infrastructure IPs. That doesn't break traceroute. Using a loopback on the other hand takes away a lot of useful information from traceroutes. For this reason I would suggest - for IPv6 at least - to use global unicast adressing on point-to-point links, as oppossed to non routable prefixes. Again that doesn't mean the /127 needs to be actually routed/redistributed. For IPv4 money will drive that decision anyway, although if I'm thinking about cross connect fees in carrier hotels I have a hard time seeing how a public /31 would be a showstopper. lukas

Yes, but this just generally reflects how poor they are run over there. Any competent sales organization would just bundle the /31 cost into the existing MRC they charge for a BGP session, which is also crap but people are paying it, so good for them I guess. That being said, this was entirely predictable. Cogent has been free cash flow negative since 2023 they bought the old Sprint assets, and that includes the gobs of money that TMobile paid them to take it. When you have to borrow money backed by revenue from IP space, **and the resale value of other IP space**, just to RUN THE BUSINESS, you have a problem. That problem gets worse when IPv4 resale value drops by 50%, as it has since then. So of course they are trying to charge whatever they can get from any IPs that get used. They are squeezing blood from every stone they can just to survive at this point. On 9/7/26 3:36 AM, Christopher Hawker via NANOG wrote:

I can either pay $400/month for a 1G transit service and $5/month for the /31 used by the BGP session, or I can pay $405/month for the transit service with a /31 free.

Either way, I’m still paying a total of $405/month. Only benefit to the former is if my upstream will allow me to use my address space for the PtP link, saving $5/month.

Regards, Christopher Hawker

Sent from my iPhone

...

On 7 Sep 2026, at 12:59 pm, Gary Sparkes via NANOG <nanog@lists.nanog.org> wrote:

The interface addresses for an IPv4 BGP session should be included in the BGP session cost.

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Monday, September 7, 2026 3:03 AM To: Vasilenko Eduard <vasilenko.eduard@huawei.com> Cc: North American Network Operators Group <nanog@lists.nanog.org>; Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

This has been raised on the thread already a few times, by different wording, for example Kevin raised the point in a bit different wording.

Everyone charges for everything already, because everyone wants cost+, how this is structured is where marketing and psychology lives.

...

On Mon, 7 Sept 2026 at 09:57, Vasilenko Eduard <vasilenko.eduard@huawei.com> wrote:

Reminder, cloud providers charge for every single IPv4 address. Why network provides should give it for free? It is not network providers have started to charge for IPv4. Ed/

...

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Friday, September 4, 2026 21:55 To: North American Network Operators Group <nanog@lists.nanog.org> Cc: Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

On Fri, 4 Sept 2026 at 21:47, Brian Knight via NANOG <nanog@lists.nanog.org> wrote:

Hey Brian,

...

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

This might be of interest: https://www.trex.fi/service/turku.html

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/ DMDSA VSOHF5GZEHMAU2ATLGLHCQAK46A/

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/G7Z4ACHN... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/B2XDYF4F...

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/DI7TUOYO...

On 7 September 2026 09:36:13 CEST, Christopher Hawker via NANOG <nanog@lists.nanog.org> wrote:

I can either pay $400/month for a 1G transit service and $5/month for the /31 used by the BGP session, or I can pay $405/month for the transit service with a /31 free.

Either way, I’m still paying a total of $405/month. Only benefit to the former is if my upstream will allow me to use my address space for the PtP link, saving $5/month.

Regards, Christopher Hawker

Sent from my iPhone

...

On 7 Sep 2026, at 12:59 pm, Gary Sparkes via NANOG <nanog@lists.nanog.org> wrote:

The interface addresses for an IPv4 BGP session should be included in the BGP session cost.

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Monday, September 7, 2026 3:03 AM To: Vasilenko Eduard <vasilenko.eduard@huawei.com> Cc: North American Network Operators Group <nanog@lists.nanog.org>; Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

This has been raised on the thread already a few times, by different wording, for example Kevin raised the point in a bit different wording.

Everyone charges for everything already, because everyone wants cost+, how this is structured is where marketing and psychology lives.

...

On Mon, 7 Sept 2026 at 09:57, Vasilenko Eduard <vasilenko.eduard@huawei.com> wrote:

Reminder, cloud providers charge for every single IPv4 address. Why network provides should give it for free? It is not network providers have started to charge for IPv4. Ed/

...

-----Original Message----- From: Saku Ytti via NANOG <nanog@lists.nanog.org> Sent: Friday, September 4, 2026 21:55 To: North American Network Operators Group <nanog@lists.nanog.org> Cc: Saku Ytti <saku@ytti.fi> Subject: Re: Charging for IPs

On Fri, 4 Sept 2026 at 21:47, Brian Knight via NANOG <nanog@lists.nanog.org> wrote:

Hey Brian,

...

From a configuration standpoint, without testing it, is it as easy as configuring an IPV6 BGP neighbor with an IPV4 unicast AFI/SAFI? And the received routes appear in the table with an IPV6 NH?

This might be of interest: https://www.trex.fi/service/turku.html

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/ DMDSA VSOHF5GZEHMAU2ATLGLHCQAK46A/

-- ++ytti _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/G7Z4ACHN... _______________________________________________ NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/B2XDYF4F...

---

NANOG mailing list https://lists.nanog.org/archives/list/nanog@lists.nanog.org/message/DI7TUOYO... Perhaps it does work like this in networking, I don't know as I don't manage our contracts, but I'd be surprised because it's not how B2B contracts work anywhere else.

The total price is the amount of money they think they can get you to pay. There are many tricks to get you to pay more. One trick is to itemize a contract. Another trick is to appear to have standard line items. You're more likely to pay an extra $5 for a BGP session if someone presents you a price list that says "BGP session: $5" than if they don't, but that list could have been created or tweaked just for your negotiation. Different customers could get a completely different price list, unless it's publicly advertised on their website. Chances are they have some internal guide prices to both estimate how much a service is worth to a customer and to make sure they don't lose money, but they have no obligation to show these to customers. Often it's assumed everyone in the negotiation already knows it works this way so they skip the line item discussion and just focus on the total price. "We can do what you want for $400" "Yep sounds good" and then the line items are made up after the fact: physical connection $250, bandwidth $130, IP rental $10, BGP session $10. And the different ways two companies may itemize their total prices has absolutely no bearing on anything whatsoever. Kevin

On Tue, 8 Sept 2026 at 01:39, Lukas Tribus <lukas@ltri.eu> wrote:

Thanks, that would be RFC 5837, I did not know about that one.

Looks like Arista has joined Juniper in implementing this [1] and Linux has it since 6.19.

[1] https://arista.my.site.com/AristaCommunity/s/question/0D5Uw000001yvSvKAI/eos...

Awesome, can't. see the actual Arista url, would be interesting in specifics. In Juniper the problem is a) basically one platform (QFX, no MX no PTX) b) TTL exceeded doesn't get the EO, making it basically useless We've opened an ER for this, but it's not a priority for Juniper. I feel like RFC5837 is such that once we get it, we wonder how we survive before it. But I seem to be fairly alone here (so wrong I guess), after many years of asking j-nsp and c-nsp to push their vendors for it. -- ++ytti

Thank you Mr. Beecher, and though I agree the size of the image is small, I do have to agree with Mr Tillery that the image is inappropriate use of NANOG, and *the time of hundreds of network operators far above my level.* After sending it, in an attempt to *be at least slightly useful* by providing some Friday afternoon levity, I did second guess sending the meme, and wished I had utilized my mother's wisdom and used my words instead. Apologies to those I ruffled, and I promise *any standard I apply to myself* in the future will be, at least, an effort to equal the esteem gathered here. michael brooks Enterprise Network Architect Adams 12 Five Star Schools michael.brooks@adams12.org :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: "flying is learning how to throw yourself at the ground and miss" On Sun, Sep 6, 2026 at 10:05 AM Tom Beecher <beecher@beecher.cc> wrote:

Whilst I may not be a fan of all the signature fufu that some people include in their emails, Mr. Brooks' entire message, inclusive of headers, is 16.2KB. Excluding headers, 2.1KB.

A fair bit away from 'most of a megabyte' I would say.

...

On 4 September 2026 21:46:42 CEST, michael brooks - ESC < michael.brooks@adams12.org> wrote:

...

[image: image.png]

michael brooks Enterprise Network Architect Adams 12 Five Star Schools michael.brooks@adams12.org :::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::: "flying is learning how to throw yourself at the ground and miss"

On Fri, Sep 4, 2026 at 1:36 PM Kevin Tillery via NANOG < nanog@lists.nanog.org> wrote:

...

I'll be the contrarian here. Surely whether or not this is sane

depends on

...

the total price and surrounding circumstances. If a provider charges say $200 for a service, whether they itemize it as $200 connection, free BGP, free IPs; or $180 connection, $20 BGP, free IPs; or $1 connection, $9 BGP, $190 for IPs to run BGP on; is entirely up to them and makes little difference. _______________________________________________ NANOG mailing list

https://urldefense.com/v3/__https://lists.nanog.org/archives/list/nanog@list...

...

-- This is a staff email account managed by Adams 12 Five Star Schools. This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender.

Is this really an appropriate use of NANOG?

Often when I post to NANOG, I am worried that I am being the "reply guy", wasting the time of hundreds of network operators far above my level with some nonsense, and I do not post things that I think will not be at least slightly useful. And I am worried that I am not skilled in

On 9/5/26 8:14 PM, Kevin Tillery via NANOG wrote: professional English writing. But this is far below any standard I apply to myself.

...

This image is now wasting most of a megabyte in how many important

people's mailboxes?

...

Kevin _______________________________________________ NANOG mailing list

https://urldefense.com/v3/__https://lists.nanog.org/archives/list/nanog@list...

-- This is a staff email account managed by Adams 12 Five Star Schools.  This email and any files transmitted with it are confidential and intended solely for the use of the individual or entity to whom they are addressed. If you have received this email in error please notify the sender.

On Thu, 10 Sept 2026 at 07:48, Owen DeLong via NANOG <nanog@lists.nanog.org> wrote:

If you’re going to ugly up the trace route that way, probably better to use Link Local (169.254.x.x) addresses than RFC-1918.

That way, the router (should) report a loopback address instead of the interface address in the trace route. Better to have a routable Loopback reported than an RFC-1918 address in most cases.

Can you tell me some example platform(s) that work like this? Any platform I've come across, if you want to send ICMP from a loop, you need to define it, and will then not care what IP you have on the interface. But I do agree that RFC1918 as a link-net on the Internet makes very little sense. -- ++ytti

On Thu, 10 Sept 2026 at 23:09, Lukas Tribus <lukas@ltri.eu> wrote:

...

If we can find actual production traceroute, we're sure to get implementations in:

https://github.com/fujiapple852/trippy https://github.com/lance0/ttl

I have filed feature requests for both with those captures.

...

At least trippy will already print all unknown EOs too, so you don't need tshark to confirm.

Yes, with -e --tui-icmp-extension-mode all trippy shows unknown extensions.

Also note that traceroute (-e) already has support for RFC5837 since 2.1.4 (so this is available in Ubuntu 24.04 LTS and Debian Trixie today): https://sourceforge.net/projects/traceroute/files/traceroute/traceroute%202.... Lukas