IVTF Discovery: The Wall wants Spaghetti thrown at it.

My day job is basically to write DHCP code. DHCPv4, DHCPv6, just DHCP in general. So one of the things I wind up doing is subscribing to the IETF DHC WG mailing list, where its volunteers work on DHCP related standards. Enter the DSL Forum. The right soundtrack to this blog entry is Re: Your Brains. You might want to go click the 'play' button in another window while you read on.

They have this problem. Some of their networks aren't as secure as they would like them to be. People can (and I guess, do) unplug their own DSL connection and swap with a neighbor who's paid for better service. Because the network is configured per-port, it works, and unless someone complains no one knows. There are a lot of solutions for securing access to "layer 2 networks". 802.1x. PANA. Just flat out physical security. But these aren't adequate, no, they need a completely new DHCP solution.

So the -00 of this draft was presented at an IETF DHC WG meeting. I was there. I remember we raised some showstopping technical and architectural concerns (packet sizes, protocol changes to permit servers to make 'requests' of INIT-state (unaddressed) clients). The author asked for the document to become a WG item anyway, the consensus was against. That's a day's work at most IETF's. But the authors of the draft, likely due to pressures from the DSL Forum, weren't satisfied with the WG consensus. It didn't agree with theirs, you see.

Chorus: "All we wanna do is...eat your brains. We're not unreasonable...I mean no one's gonna eat your eyes..."

So at the next IETF, they show up at the INT Area meeting with an agenda item of asking for the Internet Area to put the draft on the DHC WG's charter. This way you don't need DHC WG consensus (those guys don't know what they're talking about anyway), you just need INT Area goons to agree to it. The plan couldn't fail. Except it did. Quite a few of the DHC WG volunteers showed up and raised the same technical and architectural concerns, and wondered why the author had spent no time answer them. "They are not issues, you are all just in a conspiracy against me," was the vague reply. Even INT Area goons weren't about to advance that agenda.

Chorus: "All we wanna do is...eat your brains. We're at an impasse here, maybe we should compromise?"

As if this already wasn't enough of a confrontation, they continued to press the issue, appearing I guess at the last IETF DHC WG meeting, where they said...I can't imagine what. The drafts haven't changed despite all the technical problems, no architectural decision has been made obvious, and DHC hasn't posted its minutes.

They're still going on about it like it's a foregone conclusion. They've written code that implements it, they say, which is really IETF code for "even if you do not standardize it, we will just do it anyway." Somehow they've managed to get the DHC WG to review their document as if it was a WG item without reaching the consensus that it should ever be a WG item to start with.

So what's the deal? The DSL Forum is supposed to represent DSL operators. There are technologies that perform this kind of security already, as their main function. The truth is that they are more expensive than duct-taping DHCP into a role it was never designed for. Meanwhile, the duct-taped DHCP becomes much more expensive ("protocol maintenance costs") for everyone not represented by the DSL Forum. DSL operators want this?

That blank wall is just asking for it. You know you want to throw spaghetti at it. Do it.