Re: [v6ops] Combining IPv6 ND and DHCPv6 into a single, unified function

[Mikael Abrahamsson <[e-mail]>]

On Tue, 21 Nov 2017, Mark Smith wrote:

> Hi Fred,
>
> I think this fundamentally violates the principle that the network is
> application agnostic or transparent, allowing it to support any new host
> residing application without any upgrades to the network. I think the
> principle of network application transparency should apply to application
> configuration.
>
> We don't have to upgrade routers to carry new application protocols, we
> shouldn't have to upgrade routers to support configuring new applications.
>
> "Internet Transparency and Application Configuration"
>
> https://tools.ietf.org/html/draft-smith-v6ops-ce-dhcpv6-transparency-00#section-2

DHCP is a pain for rapidly changing (perhaps mobile) environments. It's 
fundamentally a polling protocol where the network hands out resources for 
a certain amount of time, and there is little chance to invalidate them. 
It's stateful.

Devices need addresses to communicate. So if environments that are dynamic 
and require fast/short setup times to gain addresses, then we need the 
network to support that.

This is not an "application" as in a program running somewhere. This is a 
deployment scenario where devices need addresses quickly with short 
turnaround times and in a bandwidth constrained environment. Saying we 
can't change the network to support this is just silly.

So you think 6lo was a bad idea and violated the network agonostic 
principle?

-- 
Mikael Abrahamsson    email: [e-mail]