Re: Machine Identity

[Miika Komu <[e-mail]>]

On Tue, 26 Feb 2008, Stephane Bortzmeyer wrote:

Hi,

> On Tue, Feb 26, 2008 at 02:11:55PM +0100,
> Jeroen Massar <[e-mail]> wrote
> a message of 47 lines which said:
>
>> IMHO HIP (RFC4423) is that exact identity layer
>
> I feel the same way but it does not seem that HIP interests many
> people?

_o/

>> The big problem though is that actually getting programs to use this
>> is cumbersome, simply because a lot of programmers don't know about
>> it.
>
> Via draft-ietf-hip-native-api? How many implementation? How could I
> use it on a machine which has no HIP in its kernel?

Yes, the native API is an explicit way of using HIP:

http://www.ietf.org/internet-drafts/draft-ietf-hip-native-api-04.txt

Fortunately, also legacy applications can also use HIP:

http://www.ietf.org/internet-drafts/draft-ietf-hip-applications-02.txt

While waiting for changes to libc, the DNS interaction can be accomplished 
by running a DNS proxy in the localhost that can do the HIP magic. The 
kernel support for the new beet mode is not mandatory because it can be 
implemented in userspace using tun/tap devices like OpenHIP does it 
already. Our project (InfraHIP) has contributed already the kernel patches 
to the Linux kernel and half of them have been accepted.

-- 
Miika Komu                                       http://www.iki.fi/miika/