Well users loose for now

The Well has two IP feeds into the internet, due to
political/biz_profit/sysadmin problems the two feeds do NOT back each
other up, and the one that can be seen from toad.com has been down for
a few days, thus blocking your email.
  If the feed comes up today, well.com based subscribers to
cypherpunks should get all of the list traffic in one big burst, if it
stays down much longer then the toad.com mailer will start trashing
the older mail (and bug us about each lost message).
 
  This points to an interesting problem, TCP/IP is very good at
getting data from point A to point B, so good that a month of bombing
by the Department of War does not destroy network conductivity.  Yet
here in the (so very rich) USA we have lots lots of outages to parts
of the net because TCP/IP does not solve the problem of PAYING for
getting packets where they belong.
  This is a protocol issue on two levels, the first is that the idea
of accounting for packets/bandwidth/capacity (in some unknown manner)
was left out of TCP/IP in the first place.  The second is that it is
going to be a very large amount of work to replace the current plant
of TCP/IP hardware and software as none of these protocols are
negotiated, just presumed to be the only way to do things.

  If we build systems like these, then they will be 'brittle' and
might inspire a few folks but not be usable by most, and some other
(large) organization will build what it wants on top of our dreams
rubble.
  There is little chance that I will ever use a monetary system that
is so brittle that it fails if one link is down or one cypher
unuseable due to it's being cracked by some unscrupulous agency.

  This important subject of protocols is the topic of the next San
Francisco Bay Area Cypherpunks meeting.

		||ugh Daniel
		Sometimes Postmaster
		hugh@toad.com

Hugh Daniel writes:

>   This is a protocol issue on two levels, the first is that the idea
> of accounting for packets/bandwidth/capacity (in some unknown manner)
> was left out of TCP/IP in the first place.  The second is that it is
> going to be a very large amount of work to replace the current plant
> of TCP/IP hardware and software as none of these protocols are
> negotiated, just presumed to be the only way to do things.
> 
>   If we build systems like these, then they will be 'brittle' and
> might inspire a few folks but not be usable by most, and some other
> (large) organization will build what it wants on top of our dreams
> rubble.
>   There is little chance that I will ever use a monetary system that
> is so brittle that it fails if one link is down or one cypher
> unuseable due to it's being cracked by some unscrupulous agency.

Indeed. "Brittleness" is what's making the creakiness of the Net all
the more apparent and critical every day. (Some things it does very
well, and I'm amazed that it works as well as it does...a lot of
clever people out there making patches.)

Software has been compared to building a Boeing 747, except that
flipping one little switch accidentally can make the wing fall off.

We see islands of relative stability (word processors, apps, etc.)
separated by flaky, error-prone (human, mostly) networks, with little
interoperability.

We mainly do "encryption" and "remailing" with our wonderful crypto
tools--and we *don't* do much of the other neat stuff that is
possible--for a simple reason: the only thing all of our myriad mail
systems, newsreaders, various platforms, and communication systems can
reliably communicate to each other is the _simple text message_!

Although much more complicated objects are in principle
intercommunicable (and Mosaic can do images, etc., so things are
changing), the basic object of communication is the text block. It can
be encrypted/decrypted, signed, and remailed, with people at the
receiving end knowing how to handle it....that's why PGP and remailers
work.

The other protocols rely more on complicated objects, signals sent
back and forth, and are much less interoperable and semantically more
ambiguous. 

>   This important subject of protocols is the topic of the next San
> Francisco Bay Area Cypherpunks meeting.

I agree. If the meeting is still on the 14th, the normal "second
Saturday," I'll be there. If it's been moved to the 21st, as was being
talked about, I'll be down in Los Angeles (where I'd still like to
meet with any LA-area Cypherpunks who wish to meet).

--Tim May



-- 
..........................................................................
Timothy C. May         | Crypto Anarchy: encryption, digital money,  
tcmay@netcom.com       | anonymous networks, digital pseudonyms, zero
408-688-5409           | knowledge, reputations, information markets, 
W.A.S.T.E.: Aptos, CA  | black markets, collapse of governments.
Higher Power: 2^859433 | Public Key: PGP and MailSafe available.
"National borders are just speed bumps on the information superhighway."

On May 4,  8:37am, Hugh Daniel wrote:
>   The Well has two IP feeds into the internet, due to
> political/biz_profit/sysadmin problems the two feeds do NOT back each
> other up, and the one that can be seen from toad.com has been down for
> a few days, thus blocking your email.

This is not quite accurate.

The WELL has one IP feed into the Internet, through BARRNET.  The WELL also
has a private T1 line to a TLG customer, and traffic between the WELL and
TLG is routed through this link.  No packets to the rest of the Internet
are sent through TLG.

At least, that was the case.  I just updated our routing tables to send
TLG traffic through BARRNET like everything else, since that TLG link is
down right now.