Re: plug: thesis on site multihoming

[Pekka Savola <pekkas@netcore.fi>]

An update,

1) yes, the thesis _is_ in English ;-)

2) the files have been updated slightly, mainly editorial updates but I've
also added the abstract and some other relatively minor changes.

For the curious, the abstract is below.

Thanks for interest.

======
Site multihoming means end-sites connecting to multiple separate network
service providers; currently, no IPv6 site multihoming
mechanism has been widely accepted yet.

This thesis studies both IPv4 and IPv6 site multihoming mechanisms using
literature, other similar studies, analysis of route advertisements at the
Finnish exchange point FICIX, and queries on multihoming practices to
major ISPs in Finland.

Currently in IPv4, there seem to be three-four main mechanisms which are
used to achieve at least some of the multihoming benefits: obtaining own
address space and AS number and advertising those, advertising more
specific routes with different path, using multi-connecting and leveraging
NAT.

In IPv6, the first two of IPv4 mechanisms which are considered
architecturally unscalable have been operationally prevented for now and
the fourth does not exist.

Based on a tentative roadmap introduced in this thesis, organizations are
split to four categories: minimal, small, large and international; each
have different multihoming requirements which can be met in different
ways.

Focusing on immediate and short term solutions, minimal organizations do
not seem to require a solution, small ones could use multi-connecting,
host-centric multihoming or multihoming at site exit routers approaches,
large ones the same or possibly separate provider independent address
allocations, and international ones either provider independent
allocations or be broken down to multiple large organizations and using
those techniques.

It is apparent that only a limited amount of work is needed to enable
sufficiently good multihoming mechanisms which should provide the required
features.  However, as the mechanisms are unarguably more difficult for
the end-site, while taking the global Internet routing architecture better
into the account, it is unclear whether they might be adopted.
======

On Fri, 28 Mar 2003, Pekka Savola wrote:

> Hi,
> 
> Sorry for a plug, but it seems relevant and may be interesting for lots of
> people.
> 
> I'm about to submit a MSc thesis on the subject of "Examining Site
> Multihoming in Finnish Networks", first looking at IPv4 and then IPv6.
> 
> If you're interested, it's available at:
> 
> http://staff.csc.fi/psavola/di.ps (.pdf is also there but quality is 
> worse)
> 
> There's some minor work to be done still (like writing the abstract,
> making a few pictures clearer and actually reading it properly myself ;-),
> but it should pretty close to done (submitting for review in the beginning
> of next week, making small modifications up until about 2 weeks from now).
> 
> Even if you don't have the time, I intend to split off a modified section
> (likely based on sections 6.1 and 6.2) and submit it as an I-D in a week
> or two.  By breaking down the sites by a few different types and looking
> at which multihoming requirements they have, we seem to be able to solve
> almost all problems at least to some extent quite nicely.  Of course, with
> long term, things can be improved a lot.
> 
> Comments etc. are of course really appreciated; off-list is probably best
> at least for the most of them.
> 
> In another w.g. I made a joke about most thesis's being non-relevant and
> useless (for the IETF).  Hopefully this doesn't fall into that category
> :-).
> 
> 

-- 
Pekka Savola                 "You each name yourselves king, yet the
Netcore Oy                    kingdom bleeds."
Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings