RE: [RRG] On the Transitionability of LISP

[András Császár (IJ/ETH) <Andras.Csaszar@ericsson.com>]

>>>              SHIM6  Six/  Mobile  LISP-   LISP-   eFIT-  Ivip
>>>                     One   IPv6    NERD    CONS    APT
>>> 
>>> Address
>>> portability                       Y       Y       Y      Y
>>> 
>>> Multihoming   Y     Y             Y       Y       Y      Y
>>> 
>>> Mobility                  Y                              Y*
>>> 
>>> IPv4 too                          Y       Y       Y      Y
>>> 
>>> No host                           Y       Y       Y      Y* changes
>>> 
>>> * Mobile IPv4 or IPv6 hosts making use of Ivip will need new host

Hi!

I think we should further grow the table.

In the above table the rows are something like goals (challanges) and conditions, and the columns are more or less newly proposed solutions.

I would find it interesting, if we could have a big table in some online collaborative spreadsheet application where we could add all other goals/conditions, and also the solutions used today for the different problems! Would there be interest in creating such a big table in collaboration? We could do it with Google Spreadsheet but that can be accessed only with a Google account. Is there some other tool? Or does everyone have a Google account? ;-)

Anyway, regarding the goals, I think we can group goals as end-user/enterprise network goals and the goals of providers (ISPs). So, here is a list for a shoot-out :-) (in bracket the solution used today)

End-user/Enterprise network goals:
- Host multi-homing (applications reconnect - no session continuity, but who cares...)
- Proper end-to-end path selection for appropriate service level per application/per flow (x)
- Avoid or easy renumbering when changing ISP (PI addresses, or PA addresses with DHCP reconfiguration)
- Site multi-homing with "proper" TE (x - of course site-multihoming for resiliency is easy)
- Mobility for the host (L2 solutions and MIP, PMIP in L3)
- Privacy of transport, protected addresses (Firewall, IPSec)

End-user/Enterprise networks' expected properties of any solution:
- No host changes
- No or easy to upgrade CE customer edge router changes
- Backward compatibility (i.e. ability to communicate) with sites/end-hosts not having the improvement
- IPv4&v6 support


Provider(ISPs,Tier-x provider) goals:
- Support of all above enterprise goals - e.g. in the form of a service
- Scalable routing (x)
- Proper Traffic Engineering (only what BGP MED and BGP tricks offer)
- Simpler management (centralised mgmt systems for remote/auto configuration?)
- Scalable VPN support (L2VPN and L3VPN solutions, most require MPLS at least in the PE router)
- Easy inter-/intradomain multicast support (? - is today's multicast good enough?)

Providers' (ISPs,Tier-x provider) expected properties of any solution:
- Easily managable 
- No or easy to upgrade and cheap provider router changes
- IPv4&v6 compatible


Router vendor goals:
- Support of above ISP goals

Router vendors' expected properties of any solution:
- Big routing table or lot of different new features ;-)


Can we extend or repair this list? And match the proposed solutions and the existing ones to each aspect?

Regards,
András
�������zǧu���Ơz��
ު笶� r������ۧ����z)ڲ)�b�欶�z����w&�r�zm���(����?~��� r����+-���