[ih] Making sense of the end-to-end arguments (or at least trying to)

Matthias Bärwolff mbaer at cs.tu-berlin.de
Wed Dec 8 03:16:04 PST 2010


Dear all,

it is my pleasure to announce the release of my PhD thesis on
"End-to-end arguments in the Internet" as a self-published book on
CreateSpace.com (https://www.createspace.com/3498028). Of course there
is a perfectly fine free PDF version on my website, too
(http://www.bärwolff.de/publications/2010-10-PhD-thesis.html).

This is also an apt opportunity to thank all you folks for the feedback
on many of my questions here on this list over the last two years. So
without further ado, I'll let you go. If you have any feedback or
questions about the book, just send me an email.

Matthias


P.S.: Some further asides for the not-too-hurried reader. The thesis is
basically split into three big parts:

- Part 1 reviews the genesis of the end-to-end arguments as a design
principle of sorts, going back to early 1960s contributions by Paul
Baran and running all the way to more timely elaborations of the merit
of minimality of common shared spanning layers. I have not given too
prominent a consideration to the notion of net neutrality; some
footnotes go into that (tracing the notion to a 1999 Saltzer note, and,
slightly more adventurously, to early 1970s Pouzin papers), but I have
not given it a section of its own.

- Part 2 then discusses the evolution the end-to-end arguments as a
functions of actual networking (as opposed to theoretical reasoning)
from the Arpanet to the Internet, along with aside notes about Alohanet,
Ethernet, etc. The contribution of this part is somewhat less tractable
than that of Part 1. At the very least, it brings a whole lot of
instructive and somewhat novel data points to the table -- e.g.
elaborating the inner structure of the Arpanet; the eventual "frazzling"
of the Arpanet edges (VDH interface, TIPs, etc.); and the history of
Arpanet raw messages as a somewhat direct predecessor to the later IP
protocol.

- Part 3 wraps up the thesis by adding some looser discussion about the
merit, useful scope, limitations, and proper articulation of the
end-to-end arguments. It somewhat defeats a two-sentence summary (and is
also strictly speaking off-topic for this list) so go and read it for
yourself.

In all, the core of the thesis runs 160 pages; plus 277 pages of
endnotes; plus 89 pages of literature; plus prelims and tables (toc,
tof); plus, last but not least, a really cool index of interesting
asides. While all of this jazz makes the most sense in the PDF version
which features all conceivable hyperlinks forth and back; the links may
also easily be followed manually (even from the endnotes and literature
back to the main text).

While the PDF is free, the book is $25, has a neat cover, and is
available on CreateSpace as of now; it will also hit Amazon.com at
around new year's eve, and eventually it will be on the international
Amazon sites some time next January. Being the author, I can order
copies of the book for some $12 and have them shipped to any US address
of my choosing; so if you think that there is a good reason for you to
have me get the book to you in some sort of out-of-band mutual deal,
drop me an email.


-- 
Matthias Bärwolff
www.bärwolff.de



More information about the Internet-history mailing list