[Battlemesh] Host Identity Protocol, any experience?
willi.uebelherr at riseup.net
Mon Nov 28 23:30:58 CET 2016
many thanks for your contribution. But in general, we should look to
that, what we need, and not to Cisco, what they think, what we need.
But this is not only valid for Cisco. Also in the IETF, if they are not
working in/for Cisco, they act for that or similar.
The basis is not a transport problem. Normally it would be easy to
continue our packet-transfer from different points. The problem is the
TCP/UDP session. This is a relict from the time, that the transport of
digital data was organised with analog systems. And with the
elctro-mechanical relais it is impossible to organise the error-checking
on the path.
Therefore, the core, or base, of MIP or HIP are the continuation of this
sessions from different points. And nothing else. Independent, what we
find in all this declarations and statements.
And in the next future? The cars are the AP. The bus are the AP. In the
train and airplanes we have it.
But this means, that this most mobile "mobile devices" never change
their IP identity. Only the routing path to the AP change.
For me, all this are huge mist clouds to distract from the real
questions. For what we need TCP/UDP in a time, where all packet
transports are digital and it is easy to check the validity?
many greetings, willi
On 28/11/2016 18:00, Tom Henderson wrote:
> On 11/28/2016 11:10 AM, willi uebelherr wrote:
>> Dear Tom,
>> we have to separate something in this discussion in the architectural
>> 1) Under some conditions, mobile devices change her AP space. And this
>> means, it can be possible to use an open session with his move.
>> This is mostly not necessary for laptops or notebooks. This can be
>> necessary for handheld, small mobile device.
>> And this mechanism is only necessary to continue an open session. If we
>> create a new session, we work with this new AP space.
>> 2) Under such specific conditions it is not necessary to inflate more
>> and more this confusion in the IP stack.
>> 3) This living sessions (TCP, UDP) was started from a specific AP with
>> the communication partner inside or outside of this AP space. And this
>> specific mobile device change the AP space. Therefore for me, it will be
>> much clearer and simpler, to conentrate to this temporal action, the
>> forwarding. Based on this focus to the AP's, a device with much more
>> resources, mostly, it will be much easier to organise the following of
>> the pakets.
>> And this is the basic architectural goal of the Mobil-IP. The
>> implementation maybe can be the same nonsense like in the HIP area.
> Hi Willi,
> If one looks generally at the two architectures (MIP, HIP), one can see
> many similarities in the approaches, from different starting points and
> emphases. MIP was originally oriented towards home-agent tunneling,
> while HIP was originally oriented more towards end-to-end handling of
> mobility. But MIP was later extended for route optimization, and HIP
> was later extended with home-agent-like functionality (HIP rendezvous
> servers, HIP relays).
> I was not advocating for HIP here, merely trying to clarify and summarize.
> - Tom
> Battlemesh mailing list
> Battlemesh at ml.ninux.org
More information about the Battlemesh