[9fans] verifying CD vs. iso image

[9fans] verifying CD vs. iso image

[pac]

Hi folks,

can you show me how to compare (checksums of) CD-R versus its original iso
image (used to be copied to /mnt/cd/wd)?
I don't know how to access the raw data on the CD  (I feel that it should be
some trick with with /dev/sdC1/ctl or  /dev/sdC1/raw ... )


Also, I've got trapped in (presumably easy to write) script to compare all
files in two (intentionally identical) directories. Just something like cmp
filelist1 filelist2. I ended up with ugly enough catting everything together
and piping it to cmp   :-(((
Beg for your help here, too. I must be an idiot, or what.


Cheers,
++pac.

[9fans] pop3 before smtp

[Kenji Arisawa]

Hello,

Does anyone have a plan to add "pop3 before smtp" to Plan9 smtpd?

Kenji Arisawa

Re: [9fans] verifying CD vs. iso image

[David Presotto]

[-- Attachment #1: Type: text/plain, Size: 44 bytes --]

How about
	sum file
	sum /dev/sdD0/data
?

[-- Attachment #2: Type: message/rfc822, Size: 2633 bytes --]

From: "pac" <cej@gli.cas.cz>
To: "9fans" <9fans@cse.psu.edu>
Subject: [9fans] verifying CD vs. iso image
Date: Mon, 7 Jul 2003 14:05:55 +0200
Message-ID: <003201c34480$1e46e090$2a8be793@gli.cas.cz>

Hi folks,

can you show me how to compare (checksums of) CD-R versus its original iso
image (used to be copied to /mnt/cd/wd)?
I don't know how to access the raw data on the CD  (I feel that it should be
some trick with with /dev/sdC1/ctl or  /dev/sdC1/raw ... )


Also, I've got trapped in (presumably easy to write) script to compare all
files in two (intentionally identical) directories. Just something like cmp
filelist1 filelist2. I ended up with ugly enough catting everything together
and piping it to cmp   :-(((
Beg for your help here, too. I must be an idiot, or what.


Cheers,
++pac.

Re: [9fans] verifying CD vs. iso image

[boyd, rounin]

err, cmp?

     http://plan9.bell-labs.com/magic/man2html?man=cmp&sect=1

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

> Does anyone have a plan to add "pop3 before smtp" to Plan9 smtpd?

I certainly hope not. RFC2554 has been out for over four years now.

--lyndon

Re: [9fans] pop3 before smtp

[boyd, rounin]

> > Does anyone have a plan to add "pop3 before smtp" to Plan9 smtpd?

i never understood that sentence.

> I certainly hope not. RFC2554 has been out for over four years now.

hmm 2554 and what might that be?

Network Working Group                                           J. Myers
Request for Comments: 2554                       Netscape Communications
Category: Standards Track                                     March 1999


                         SMTP Service Extension
                           for Authentication

Status of this Memo

   This document specifies an Internet standards track protocol for the
   Internet community, and requests discussion and suggestions for
   improvements.  Please refer to the current edition of the "Internet
   Official Protocol Standards" (STD 1) for the standardization state
   and status of this protocol.  Distribution of this memo is unlimited.

----

what has it got to do with POP?

Re: [9fans] pop3 before smtp

[Dan Cross]

> what has it got to do with POP?

Early on in the days of the spam epidemic, as a crude form of
authentication, someone modified sendmail to require a user to
authenticate to their POP server before allowing them to relay mail
through their server.  Basically, the POP server put a timestamp
associated with a user's email address in a file somewhere, and then
the MTA looked in it before relaying mail from that address.  Like most
bad ideas, it stuck.

	- Dan C.

Re: [9fans] pop3 before smtp

[Scott Schwartz]

> what has it got to do with POP?

The idea is that there's no good way to distinguish SMTP message
submission by authorized users from message relaying by unauthorized ones.
But POP does authenticate you as a local user, and it can remember
your IP address, and your mail server can assume that attempts at mail
relaying from that IP are actually attempts at mail submission and that
they should be allowed.  It's a popular kludge.  And you don't need it
if your SMTP server is willing to authenticate local users.

Re: [9fans] pop3 before smtp

[boyd, rounin]

> It's a popular kludge.  And you don't need it
> if your SMTP server is willing to authenticate local users.

oh i see.  thanks for the clarification, but the whole thing is nonsense.

as i said before:

     good crypto is never OT

clear text tcp auth needs to be replaced by something like styx.
i'll send my pop password in the clear to the server 'cos what have
i have got to lose?  coupla bits of random mail that get pulled off
the pop server every few minutes.

it's time for a re-think, no?

i am unsure how to do this, but it's gotta be a lot more interesting than
bitching
about how shell metacharacter expansion doesn't/does work on the other
side of an i/o redirection [modulo your acid flashback of 'the shell'].

Re: [9fans] pop3 before smtp

[boyd, rounin]

> Early on

any given day

> someone modified sendmail to

do an arbitrarily stupid thing.

vix had it together, back in early '92 (or so) and then it all just went to
hell.

bit like the 'new' lunix DNS ...

Re: [9fans] pop3 before smtp

[Kenji Arisawa]

Hello,


>> It's a popular kludge.  And you don't need it
>> if your SMTP server is willing to authenticate local users.
>
> oh i see.  thanks for the clarification, but the whole thing is
> nonsense.

Hmm..
How do you send emails from notebook via Plan 9 server?
"POP3 before smtp" is a simple solution though not a best one.

Kenji Arisawa

Re: [9fans] pop3 before smtp

[Fco.J.Ballesteros, nemo]

> Hmm..
> How do you send emails from notebook via Plan 9 server?
> "POP3 before smtp" is a simple solution though not a best one.

The notebook can send them by itself.

Re: [9fans] pop3 before smtp

[Kenji Arisawa]

Hello,

>> Hmm..
>> How do you send emails from notebook via Plan 9 server?
>> "POP3 before smtp" is a simple solution though not a best one.
>
> The notebook can send them by itself.

IP address of my notebook is determined by DHCP.
The address changes from place to place.

Kenji Arisawa

Re: [9fans] pop3 before smtp

[Geoff Collyer]

Your IP address shouldn't matter if you've set "site" to your domain
in termrc and cpurc, and set "fd" to your domain (if the return
address would be unqualified) in /mail/lib/remotemail*.  Various
anti-spam measures at other sites may interfere, but I don't yet
control my reverse DNS map and have had very little trouble sending
mail directly even though the domain returned by reverse lookup of my
IP addresses is different than the one I claim in mail.  In those few
cases, I've rerouted outgoing mail in rewrite to my ISP's mail
servers.  Currently the only sites my rewrite does this for are at my
old employer, the University of Toronto.

Re: [9fans] pop3 before smtp

[Kenji Arisawa]

Hello,

Thanks all.
It seems current smtpd support some relay authentications.

	SMTP authentication by login and cram-md5 protocols
           is supported; authenticated connections are permitted to
           relay.

I didn't aware that. I will try.

Kenji Arisawa

Re: [9fans] pop3 before smtp

[Kenji Arisawa]

Hello,

	SMTP authentication by login and cram-md5 protocols
           is supported; authenticated connections are permitted to
           relay.

OK. That works fine. Thanks for the person who wrote smtpd.

Kenji Arisawa

Re: [9fans] pop3 before smtp

[Russ Cox]

> How do you send emails from notebook via Plan 9 server?
> "POP3 before smtp" is a simple solution though not a best one.

I use exactly that, via ratfs(4) and pop3(8).
I'm surprised no one else mentioned this yet.
Imap4 before smtp also works.

Russ

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

On Wednesday, July 9, 2003, at 10:02  PM, Russ Cox wrote:

>> How do you send emails from notebook via Plan 9 server?
>> "POP3 before smtp" is a simple solution though not a best one.
>
> I use exactly that, via ratfs(4) and pop3(8).
> I'm surprised no one else mentioned this yet.
> Imap4 before smtp also works.

There is some invisible barrier out there that apparently stops e-mail
client authors from implementing SMTP AUTH (or STARTTLS), but I'll be
damned if I can figure out what it is. Given that most modern e-mail
clients already support SSL/TLS and SASL for POP and IMAP, what's so
difficult about making the SMTP/Submission bits authentication aware?
(Seriously. I've been trying to get a straight answer for this out of
client authors for years, without success.)

--lyndon

Re: [9fans] pop3 before smtp

[David Presotto]

as a client author, I pierced the invisible barrier by letting dan
cross write the code and stick it in.

Re: [9fans] pop3 before smtp

[George Michaelson]

probably the same twisted logic which prevented people implementing 'turn
please' in the same protocol. oh hang on, that WAS a security risk, wasn't it...

I believe it was Charles Forsyth who pointed out to me S(imple) MTP didn't make
much sense when the paper printout of the RFC was 1/2 an inch thick.

in marketing terms, I suspect embedded sender security would be popular. I use a
mailer called 'sylpheed' which is a GUI frontend to MH, and it does  STARTTLS in
its choices, and does it even if you don't want it, if the upstream SMTP server
demands it. So maybe there is some sanity emerging.

-George

Re: [9fans] pop3 before smtp

[David Presotto]

Seriously, the biggest barrier is the thing noone seems to have,
a client certificate.

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

On Thursday, July 10, 2003, at 01:33  PM, George Michaelson wrote:
> I believe it was Charles Forsyth who pointed out to me S(imple) MTP
> didn't make
> much sense when the paper printout of the RFC was 1/2 an inch thick.

I will forgo *that* discussion today, since I have to get some
productive work done. However ... if you've managed to implement that
.5 inch document, the incremental work to add AUTH with CRAM-MD5 (and
LOGIN under SSL) can't be that bloody difficult.

Okay, one e-mail protocols cheap shot: IMAP puts the S in SMTP ;-)

--lyndon

Re: [9fans] pop3 before smtp

[boyd, rounin]

the real problem is that the PKI is totally flawed.

it's a sad state of affairs when Notwork Solutions (sic), who are
pretty tight with VeriSad (sic), operate with an X.509 certificate
that's expired ...

time for a rethink, methinks.

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

On Thursday, July 10, 2003, at 01:38  PM, David Presotto wrote:

> Seriously, the biggest barrier is the thing noone seems to have,
> a client certificate.

 From a practical standpoint, you don't need one. You can use SSL to
encrypt the protocol session (without a cert), and use AUTH LOGIN
inside of that to authenticate the submitting agent.

If you don't need the encrypted link, just AUTH using a non-plaintext
SASL mechanism (e.g., CRAM-MD5, DIGEST-MD5).

--lyndon

Re: [9fans] pop3 before smtp

[boyd, rounin]

>  From a practical standpoint, you don't need one.

you do.  alice doesn't trust bob and bob doesn't trust alice.

if you're gonna do it, do it right, or just forget about it.

i am certain that some POP3/SMTP auth combination will
just break.  in the limited case such a hideous kludge will
probably work.  in the general case, i think not.

Re: [9fans] pop3 before smtp

[William Ahern]

On Thu, Jul 10, 2003 at 09:49:43PM +0200, boyd, rounin wrote:
> the real problem is that the PKI is totally flawed.
>
> it's a sad state of affairs when Notwork Solutions (sic), who are
> pretty tight with VeriSad (sic), operate with an X.509 certificate
> that's expired ...
>
> time for a rethink, methinks.

What is needed is a distributed PKI. If I can ever find this elusive race
condition in AnonNet (I hate threads; bad design decision) I would really
like to work on AuthNet, using a design in this vein:

	http://theory.lcs.mit.edu/~cis/cis-threshold.html

I'm not an expert in this area by any measure.

	"A colleague once told me that the world was full of bad security
	systems designed by people who read 'Applied Cryptography'"

	-- Bruce Schneier

But *somebody* has to fill this gap. Maybe I can scare someone enough to
take my place.

If anybody has some [debugging] time on their hands, c/o:

	http://www.authnet.org/anonnet/

- Bill

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

On Thursday, July 10, 2003, at 02:01  PM, boyd, rounin wrote:

>>  From a practical standpoint, you don't need one.
>
> you do.  alice doesn't trust bob and bob doesn't trust alice.

You can put an SSL cert on the server if you really want to verify it's
who it claims to be. (Most people don't seem to care.) You don't need
one for the client, though, as you're verifying it via SASL. (And some
SASL mechanisms can also verify the server side while authenticating
the client, as well as performing encryption of the session. You pick
the one that matches your requirements.)

Again, I'm speaking from a *practical* standpoint, and addressing a
*very* narrow problem space. I, too, would like to see a global PKI
that actually works.

--lyndon

Re: [9fans] pop3 before smtp

[Scott Schwartz]

| What is needed is a distributed PKI.

But why?  It seems easy enough to use use private keys, and a nice
protocol like SRP.

I guess the reason is that you just can't convince people not to give
their password away (just type it here in this web page!!), and, worse,
can't convince programmers not to ask for it.

Re: [9fans] pop3 before smtp

[Dan Cross]

> | What is needed is a distributed PKI.
>
> But why?  It seems easy enough to use use private keys, and a nice
> protocol like SRP.

Well, the typical reason given is that you end up with this n^2 key
distribution problem.  PKI (in theory, at least) solves that via
signature chains.  Shared secret key systems like Kerberos have
attempted to solve this with authentication hierarchies, but while
e.g.  Kerberos has proliferated, the hierarchial authentication
component hasn't.

I don't understand this talk of `distributed PKI' though; isn't the
whole idea of a PKI that it's distributed to begin with?  Supposedly we
have that; it's just never really worked all that well.

It's a shame.  Public key cryptography involves some absolutely
beautiful mathematics.  Too bad people are disgusted with it due to the
poor implementations they most frequently encounter.

	- Dan C.

Re: [9fans] pop3 before smtp

[boyd, rounin]

> It's a shame.  Public key cryptography involves some absolutely
> beautiful mathematics.  Too bad people are disgusted with it due to the
> poor implementations they most frequently encounter.

i'm with you, captain.

Re: [9fans] pop3 before smtp

[Geoff Collyer]

Yeah, SMTP is a misnomer.  It makes me wonder how complex MTP was.  I
think a lot of the complexity is due to typical Internet
micromanagement of a baroque syntax.

I designed a replacement, the Really Simple Mail Transfer Protocol,
and did an implementation or two in Limbo.  The spec. is 13 pages,
including index, rather than the 68 pages of RFC 821.  Lucent has had
a patent application in the works for years (4 or 5, I think), so I'm
not sure how much I can say about it, except to point interested
parties at (this is one line):

http://appft1.uspto.gov/netacgi/nph-Parser?Sect1=PTO2&Sect2=HITOFF&p=1&u=%2Fnetahtml%2FPTO%2Fsearch-bool.html&r=2&f=G&l=50&co1=AND&d=PG01&s1=collyer.IN.&OS=IN/collyer&RS=IN/collyer

The description therein is a legalistic version of a lawyer's
understanding of my paper, so it's likely pretty difficult to follow,
alas.

Of course, as I suggest at the end of the RSMTP paper, protocols like
SMTP and RSMTP could be replaced by 9P2000 if it were used more
universally.  (Mount the recipient's synthetic inbox and write to it.
Done.)  The last thing the Internet needs is more protocols, but bad
ones like SMTP ought to be replaced.

Re: [9fans] pop3 before smtp

[boyd, rounin]

> Yeah, SMTP is a misnomer.

nah, i think strict SMTP is fine.  it's 822 that really screws things up.

ESMTP is a debacle and the MINE (sic) field is a disaster.

i asked gettys about some of it, 'cos by the end of the 4th MINE (sic)
doc [some 150 a4 pages] i was beginning to lose it.

like francis ford coppola said about the making of apocalypse now:

    we had too much money, too much equipment and slowly, but
    surely we went insane

just like the war.

Re: [9fans] pop3 before smtp

[Lyndon Nerenberg]

On Thursday, July 10, 2003, at 04:02  PM, Geoff Collyer wrote:

> Yeah, SMTP is a misnomer.  It makes me wonder how complex MTP was.

It was FTP :-P  See RFC 221.

--lyndon

Re: [9fans] pop3 before smtp

[Geoff Collyer]

There are lots of things wrong with strict SMTP, such as the enforced
slowness due to its lock-step nature, but perhaps the most egregious
is that it enforces a 7-bit channel.  I used to think that this was
for the sake of long-dead PDP-10s, but it turns out to be for the
benefit of (CSnet?) people who wanted to run SMTP over plain old
un-error-corrected, un-flow-controlled serial lines without benefit of
TCP/IP underneath, and who feared that the high bit might be lost as a
parity bit (never mind that whole chunks of the message might be lost
due to lack of flow control, or garbled due to lack of error
correction).

So given that one expects there to be people who can only talk strict
821 SMTP, Content Transfers Encodings become necessary if one wants to
send 8-bit data, such as UTF-8.  So some of the MIME mess is directly
due to the failings of strict SMTP.  And that's just a bit of the
fall-out from SMTP's failings.

Re: [9fans] pop3 before smtp

[boyd, rounin]

i thought the whole 7 bit deal was due to the IMP's.

the number of times i've gotten into shitfights over:

    it is a 7 bit transport -- period

i had a simpler vision of ESMTP, ages ago; you just say with HELO
or something to say 'will you talk 8 bit'?  if the answer is 'no way'
you return the mail.  this way it would track the braindamage and
force the 7-bitters to go to 8 bits.  no 4 MIME docs, no zillion
lines of code.

i suggested it to vix, but he wasn't convinced.

Re: [9fans] pop3 before smtp

[Geoff Collyer]

I doubt that the 7-bit channel was due to the IMPs; FTP can use all 8
bits.  (IMP stood for Interface Message Processor: Honeywell 516 [my
memory says 316, but BBN's history disagrees] minicomputers used as
front-ends to hosts to handle the complexity of dealing with the
network.  Only the IMPs attached directly to the ARPAnet originally.)

But the 7-bit stupidity infected protocols inspired by SMTP, such as
NNTP (which protocol wasn't even necessary, but that's another story).

Re: [9fans] pop3 before smtp

[boyd, rounin]

> I doubt that the 7-bit channel was due to the IMPs; FTP can use all 8
> bits.

yeah, i guess it was 7 bit control and with FTP 8 bit data.

but all this is ancient history.  we gotta replace all that stuff.

although i don't see it happening soon :(

hell, they can bearly keep up even with the 'army of programmers'.

Re: [9fans] pop3 before smtp

[David Presotto]

[-- Attachment #1: Type: text/plain, Size: 555 bytes --]

Actually, compared to most of the standards out there, I find mime
refreshingly simple.  It actually solves a problem, albeit
with more mechanism than I would like.  I don't want to go back
to the old days of n different ways to encode cruft into a
mail message.  I also find it quite easy to present the mime
encoding as a pretty simple hierarchical file system.  I don't
really need anything to replace MIME since the only thing I'ld
be getting is a bit more simplicity and little else.

If I were to throw stones at anything it would be ASN.1.

[-- Attachment #2: Type: message/rfc822, Size: 2499 bytes --]

From: "boyd, rounin" <boyd@insultant.net>
To: <9fans@cse.psu.edu>
Subject: Re: [9fans] pop3 before smtp
Date: Fri, 11 Jul 2003 01:48:30 +0200
Message-ID: <01a101c3473d$c48a9720$b9844051@insultant.net>

> I doubt that the 7-bit channel was due to the IMPs; FTP can use all 8
> bits.

yeah, i guess it was 7 bit control and with FTP 8 bit data.

but all this is ancient history.  we gotta replace all that stuff.

although i don't see it happening soon :(

hell, they can bearly keep up even with the 'army of programmers'.

Re: [9fans] pop3 before smtp

[David Presotto]

[-- Attachment #1: Type: text/plain, Size: 809 bytes --]

The x.509 PKI is distributed.  The only problem with it is that noone
takes responsibility for anything.  I'ld be happy to believe a CA about
a cert it signs if the cert itself didn't contain a scree about how
the CA absolves itself of all responsibility.

Also, the revocation problem has also never been well handled.  There are
revocation servers out there but they tend to be overloaded and often
unavailable.  Shorter term keys would help accept then one of the big
advantages of having a CA would disappear, i.e., that you wouldn't
have to talk to it very often.  With revocation lists and/or short
lived keys, shared keys don't look so bad anymore.

The result is that the one org willing to take financial responsibility,
i.e. Microsoft, can get away with something as crufty as Passport.

[-- Attachment #2: Type: message/rfc822, Size: 2972 bytes --]

From: Dan Cross <cross@math.psu.edu>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] pop3 before smtp
Date: Thu, 10 Jul 2003 17:50:47 -0400
Message-ID: <200307102150.h6ALol704789@augusta.math.psu.edu>

> | What is needed is a distributed PKI.
>
> But why?  It seems easy enough to use use private keys, and a nice
> protocol like SRP.

Well, the typical reason given is that you end up with this n^2 key
distribution problem.  PKI (in theory, at least) solves that via
signature chains.  Shared secret key systems like Kerberos have
attempted to solve this with authentication hierarchies, but while
e.g.  Kerberos has proliferated, the hierarchial authentication
component hasn't.

I don't understand this talk of `distributed PKI' though; isn't the
whole idea of a PKI that it's distributed to begin with?  Supposedly we
have that; it's just never really worked all that well.

It's a shame.  Public key cryptography involves some absolutely
beautiful mathematics.  Too bad people are disgusted with it due to the
poor implementations they most frequently encounter.

	- Dan C.

Re: [9fans] pop3 before smtp

[boyd, rounin]

> If I were to throw stones at anything it would be ASN.1.

yes that thing is truly frightening.  could it be _more_ complex?

Re: [9fans] pop3 before smtp

[boyd, rounin]

at some point when the bank [french] were talking about doing
X.509 stuff and random things with their clients i suggested
they stuck the certs on a chip/smart card (the things being
rampant in france since a govt decree in 1991) and jamming
it into a reader.  when it dies it's dead and then you use some
other channel to renew it.

it's quasi private key, but not quite.  RSA is a good idea.  another
guy i sort of worked with told me back in '98 that you could get
quasi decent DES out of a chip card and RSA wasn't that far off.

Re: [9fans] pop3 before smtp

[David Presotto]

[-- Attachment #1: Type: text/plain, Size: 67 bytes --]

Atmel makes a reasonable chip and its also available in card
form.

[-- Attachment #2: Type: message/rfc822, Size: 2726 bytes --]

From: "boyd, rounin" <boyd@insultant.net>
To: <9fans@cse.psu.edu>
Subject: Re: [9fans] pop3 before smtp
Date: Fri, 11 Jul 2003 02:09:15 +0200
Message-ID: <01ed01c34740$aa416f80$b9844051@insultant.net>

at some point when the bank [french] were talking about doing
X.509 stuff and random things with their clients i suggested
they stuck the certs on a chip/smart card (the things being
rampant in france since a govt decree in 1991) and jamming
it into a reader.  when it dies it's dead and then you use some
other channel to renew it.

it's quasi private key, but not quite.  RSA is a good idea.  another
guy i sort of worked with told me back in '98 that you could get
quasi decent DES out of a chip card and RSA wasn't that far off.

Re: [9fans] pop3 before smtp

[bs]

Scott Schwartz wrote:
> | What is needed is a distributed PKI.
>
> But why?  It seems easy enough to use use private keys, and a nice
> protocol like SRP.
>
I think SRP or any other Zero Knowledge Proof system is the way to go.
Users can still enter passwords, as they used to, and crackers will find
very little to crack.

Re: [9fans] pop3 before smtp

[boyd, rounin]

> ... Zero Knowledge Proof system is the way to go.

yeah, that sounds quite viable.

Re: [9fans] pop3 before smtp

[matt]

>
>
>The result is that the one org willing to take financial responsibility,
>i.e. Microsoft, can get away with something as crufty as Passport.
>

There's a huge pot of gold for someone who starts this off.
Sadly it's going to take a world player to force it.

Someone with a big set of electronic customers needs to say :
"we will no longer accept SMTP mail"


The Register has this recent [scare] story :

The Internet Research Task Force's (IRTF) quest for an
effective solution to spam has struck gold in the form of
IT specialist Mark McCarron.



http://theregister.co.uk/content/55/31638.html

Re: [9fans] pop3 before smtp

[William Ahern]

On Thu, Jul 10, 2003 at 05:50:47PM -0400, Dan Cross wrote:
> > | What is needed is a distributed PKI.
> >
> > But why?  It seems easy enough to use use private keys, and a nice
> > protocol like SRP.
>
> Well, the typical reason given is that you end up with this n^2 key
> distribution problem.  PKI (in theory, at least) solves that via
> signature chains.  Shared secret key systems like Kerberos have
> attempted to solve this with authentication hierarchies, but while
> e.g.  Kerberos has proliferated, the hierarchial authentication
> component hasn't.
>
> I don't understand this talk of `distributed PKI' though; isn't the
> whole idea of a PKI that it's distributed to begin with?  Supposedly we
> have that; it's just never really worked all that well.

Because for many things, especially when you get into generic web services,
you don't need a hierarchy of _trusted_ certificate chains that you can
trace. All you really care is that the same client who visited you yesterday
is the same one doing a follow-up today. Or maybe that you were redirected
to service XYZ, and you need a high degree (not absolute) of probability
that the service XYZ you are talking to is the one you were meant to
be redirected to.

Not to mention its pretty much requisite to build any significantly sized
trust metric system.

If I'm in a corporation, then a hierarchical system is normative. But
in the rest of the world, why do I care if some capriciously chosen
entity vouches for the _name_ (not identity) of some web site?

- Bill

Re: [9fans] pop3 before smtp

[William Ahern]

On Fri, Jul 11, 2003 at 02:09:15AM +0200, boyd, rounin wrote:
> at some point when the bank [french] were talking about doing
> X.509 stuff and random things with their clients i suggested
> they stuck the certs on a chip/smart card (the things being
> rampant in france since a govt decree in 1991) and jamming
> it into a reader.  when it dies it's dead and then you use some
> other channel to renew it.
>

I bought a 10-pack of Schlumberger cryptocards (RSA operations computed
on-chip, plus the USB controller is on-chip so there's no need to carry
around a card reader everywhere). I've been meaning to setup a completely
password-less system for login to my personal machines, as well as
authentication to my servers via ssh. Now I'm writing an Apache module to
interface w/ BSD Auth (similar to PAM), so it can all integrate w/ the web
sites as well.

The only problem is I can't get the damn thing to get recognized in
Linux.... Supposedly everybody and their uncle has gotten it to work in a
snap. *sigh*

Interactive password systems stink, but like many other sticky subjects,
where's the alternative? (tho in an all Windows environment I've read its
fairly workable).

Re: [9fans] pop3 before smtp

[Martin Harriss]

William Ahern wrote:
>
> On Fri, Jul 11, 2003 at 02:09:15AM +0200, boyd, rounin wrote:
> > at some point when the bank [french] were talking about doing
> > X.509 stuff and random things with their clients i suggested
> > they stuck the certs on a chip/smart card (the things being
> > rampant in france since a govt decree in 1991) and jamming
> > it into a reader.  when it dies it's dead and then you use some
> > other channel to renew it.
> >
>
> I bought a 10-pack of Schlumberger cryptocards (RSA operations computed
> on-chip, plus the USB controller is on-chip so there's no need to carry
> around a card reader everywhere). I've been meaning to setup a completely
> password-less system for login to my personal machines, as well as
> authentication to my servers via ssh. Now I'm writing an Apache module to
> interface w/ BSD Auth (similar to PAM), so it can all integrate w/ the web
> sites as well.
>
> The only problem is I can't get the damn thing to get recognized in
> Linux.... Supposedly everybody and their uncle has gotten it to work in a
> snap. *sigh*
>
> Interactive password systems stink, but like many other sticky subjects,
> where's the alternative? (tho in an all Windows environment I've read its
> fairly workable).

But now these cards become bearer instruments.  You steal the card, you
have access.  Methinks you need at least a PIN to validate the card.

Martin

Re: [9fans] pop3 before smtp

[William Ahern]

On Fri, Jul 11, 2003 at 11:16:03AM -0400, Martin Harriss wrote:
<snip>
> But now these cards become bearer instruments.  You steal the card, you
> have access.  Methinks you need at least a PIN to validate the card.
>
> Martin

yes, but you *know* when you're card is gone. often there isn't even a
hint that somebody has _stolen_ your password.

i certainly don't think these cards are a panacea. you can strip the casing
w/ acid and trace the on-board chips and _steal_ the private key. but for
the forseeable future this behavior is significantly mediated by time and
physical constraints, which at the very least can give me a warning that the
card cannot be trusted anymore.

i would personally frown on any singular dependence by a government on the
technology, because of your point, and also the corollary that the more
valuable they become, the quicker the bad guys will develop the ability to
copy/replace them.

Re: [9fans] pop3 before smtp

[Bruce Ellis]

> > I bought a 10-pack of Schlumberger cryptocards (RSA operations computed
> > on-chip, plus the USB controller is on-chip so there's no need to carry
> > around a card reader everywhere).

I have some of these - they are kinda neat.  They generate and store RSA key pairs
and won't release the private key.

> > The only problem is I can't get the damn thing to get recognized in
> > Linux.... Supposedly everybody and their uncle has gotten it to work in a
> > snap. *sigh*

Not a good space to be looking at Linux code - there is so much
Linux smart card code - buckets of repeated code, all almost right.

> But now these cards become bearer instruments.  You steal the card, you
> have access.  Methinks you need at least a PIN to validate the card.

The cards I have are PIN protected.

Re: [9fans] pop3 before smtp

[C H Forsyth]

[-- Attachment #1: Type: text/plain, Size: 149 bytes --]

what makes you think that PKI is necessarily or even essentially hierarchical?
are you sure that was that the aim of some (or all) of the inventors?

[-- Attachment #2: Type: message/rfc822, Size: 4087 bytes --]

From: William Ahern <william@25thandClement.com>
To: 9fans@cse.psu.edu
Subject: Re: [9fans] pop3 before smtp
Date: Fri, 11 Jul 2003 07:46:43 -0700
Message-ID: <20030711144643.GA26212@wilbur.25thandClement.com>

On Thu, Jul 10, 2003 at 05:50:47PM -0400, Dan Cross wrote:
> > | What is needed is a distributed PKI.
> >
> > But why?  It seems easy enough to use use private keys, and a nice
> > protocol like SRP.
>
> Well, the typical reason given is that you end up with this n^2 key
> distribution problem.  PKI (in theory, at least) solves that via
> signature chains.  Shared secret key systems like Kerberos have
> attempted to solve this with authentication hierarchies, but while
> e.g.  Kerberos has proliferated, the hierarchial authentication
> component hasn't.
>
> I don't understand this talk of `distributed PKI' though; isn't the
> whole idea of a PKI that it's distributed to begin with?  Supposedly we
> have that; it's just never really worked all that well.

Because for many things, especially when you get into generic web services,
you don't need a hierarchy of _trusted_ certificate chains that you can
trace. All you really care is that the same client who visited you yesterday
is the same one doing a follow-up today. Or maybe that you were redirected
to service XYZ, and you need a high degree (not absolute) of probability
that the service XYZ you are talking to is the one you were meant to
be redirected to.

Not to mention its pretty much requisite to build any significantly sized
trust metric system.

If I'm in a corporation, then a hierarchical system is normative. But
in the rest of the world, why do I care if some capriciously chosen
entity vouches for the _name_ (not identity) of some web site?

- Bill