[p2p-research] Google gets into the DNS business

[Eugen Leitl]

On Sun, Dec 06, 2009 at 10:43:54AM -0500, Paul D. Fernhout wrote:

> This may reflect a deeper shift in our society. For most people, the Google 
> corporation is now effectively the de-facto government that structures 
> their lives online. So, how can we make Google a good government? :-)

You can't. Neither a corporate nor a government entity can be trusted.
You can trust other end users a lot more if safeguarded against
snooping and tampering precisely because their capabilities do
not differ markedly from your own. Cryptography is a key way to achieve that.
Cryptography is also a key technology to keep track of trust and reputation.
 
> Well, one could also put a willingness to use Google for convenience or 
> free-of-direct cost down to "digital fatalism" in just assuming you have no 
> privacy online.

I find declaring defeat a priori both unwise and dangerous. Also for other people,
not just you. Especially since you cannot solve social/political problems 
technically. 
 
> Here is how I see it, probably having an even higher level of paranoia than 

Paranoia is the wrong term, since we do not have to guess under
which level of scrutiny we are living. We know. It's mostly a matter of
public record. Very little if at all has to be extrapolated to fill
in the blanks. 

> you. :-) Vitamins like D3 and B12 can help supposedly. :-)
>   http://www.brainnutritionfacts.com/tag/paranoia
> "We recently had a patient at the hospital where I work who became 
> psychotic, or lost touch with reality, believing that the CIA wanted to 
> assasinate him, and hearing voices whispering death threats through the 
> hallways.  The only cause that could be found was vitamin B12 deficiency. 
> Once we gave him injections of vitamin B12, he slowly improved and could 
> once again tell the difference between reality and what only existed in his 
> mind."
> 
> So, here is some extreme paranoia for your discomforting. :-)

The first time I've run into security problems in regards to military
cryptography and was disciplined for it was 1988. Since then, I very
much doubt my profile has gotten any less conspicuous. Or so a little
bird told me a few years ago. 
 
> The fact is, any ISP can log and decrypt all the data anyone sends through 
> it from a specific location if they really want to. One can assume the 

Log yes, decrypt no. TLAs can do a lot more, but they have their limitations.

> decryption can be done now either at high cost by supercomputers or in a 

Nobody does brute-force decryption of modern cryptosystems. Brute-force
attack on the passphrase, yes. Sigint and system compromise, yes. Brute-force
the cryptosystem, no.

> few years at low cost by newer computers -- maybe even quantum ones. That 
> assumes the encryption algorithms are any good and don't have backdoors 
> making it trivial to decrypt stuff. Historically, countries have kept 

Regardless of potential weaknesses, if the only alternative is cleartext
whether to do or not to do encryption is a no-brainer.

> copies of intercepted messages from other countries in wartime and maybe 
> years later could then decrypt them and find out a lot about the past that 
> may be useful in the present day. Maybe what you outline with VPN might 
> keep stuff private now, but would it keep it private ten years from now 
> with decryption computers 1000 times faster or cheaper (Moore's law) than 
> what we have now?

Not my threat model. You're wrong about systems getting faster, this is 
not what cooks your goose. Without advances in cryptoattacks faster hardware
is meaningless. 
 
> Let's say someone unintentionally but illegally downloads just *one* 
> copyrighted song now (2009) using VPN technology (say, a background tune on 

Hey, they can log those few TByte/month that passes through my system.
Be my guest, I have no idea what's in there. Please store it, please burn
resources on it, that's precisely the point. Sufficiently widespread
nonclear traffic is indistinguishable from lots of honeypots.

> a youtube video they look at for other reasons), out of all their other 
> endless legal activities. Ten years later, in 2019, that logged data gets 
> decrypted routinely, and the government knows that person broke a copyright 

You have a strange model on how cryptoanalysis advances.

> law ten years ago. Well, is that a prosecutable crime in 2019?

Let's say you're organizing a strike or a demonstration today. Of what
interest to you is that somebody *might* (I consider it unlikely, session
keys are ephemeral) break a given OpenVPN session log two decades hence?

>   "Legislative history -- Copyright felony act"
>   http://www.cybercrime.gov/CFAleghist.htm
> It may depend on the interpretation of the "statute of limitations" as well 
> as other laws passed over the next decade. But, even if the statute of 
> limitations has passed, evidence of a past crime might seem like at least 
> "probable cause" to investigate further in the then present day (2020) by 
> seizing all of that person's computing systems. That might even be painful 

Use encryption today to make sure this doesn't happen tomorrow.

> for a cyborg, to rip out any brain implants. So, even today's laws can all 
> become a way of doing selective enforcement against people in the future if 

There's no guessing about that. It is routine today. You're right now in
violation of probably a dozen laws at least, and can be persecuted for it.

> the government does not like (even to the point of doing forcible brain 
> surgery on them). This sort of issue outlined in this slashdot article may 
> become all too common as law enforcement can look back on data recorded ten 
> years ago and decrypt it:
> ""Accidental" Download Sending 22-Year-Old Man To Prison"
> http://yro.slashdot.org/story/09/12/05/1511258/Accidental-Download-Sending-22-Year-Old-Man-To-Prison

I'm missing the part where they traced his Tor session and cracked his 
Truecrypt system partition which was secured by a strong passphrase.

Oh wait, because that never happened. He unwittingly ran into a honeypot, 
and submitted to have his unhardened machine to be analyzed forensically, 
then pleaded guilty in a fishy bargain in a modern equivalent of a Malleus 
maleficarum case.
 
> After September 11, 2001, examining cell phone calls was a big thing -- 
> showing how they are routinely recorded (a fact not widely know at the 
> time, but obvious after).

Much more interestingly is that in many countries service providers
must log your location info with time (which is a movement profile)
and record connection info. Plans to log your complete online trail
are in the works, at least in UK. Plans to only allow going online 
when authenticated with personal ID smartcards ditto. One wonders 
what plans are in the drawers we haven't heard about yet.
 
> Even now people are working to crack GSM phone encryption. Related:

GSM "encryption" was deliberately designed to be weak. Realtime
GSM cracks are over a decade old now. Try the same with another
cryptosystem considered deprecated, such as 3DES. Not so easy now, eh.

> "Open Source Attempt To Crack GSM Encryption"
> http://it.slashdot.org/story/09/12/05/1958251/Open-Source-Attempt-To-Crack-GSM-Encryption
> "The intended approach is to create an open source project to spread the 
> computation of a giant look-up table across more than 80 machines. 
> Interestingly, they've openly stated that nVidia's CUDA technology will be 
> used to execute parallel elements of the problem on GPUs as well.""
> 
> So, for anyone who archives all encrypted phone communications around them 
> (would take a lot of storage space obviously), they would be able in ten 
> years or so probably to listen to everything anyone around them said now on 

Not my SIP session over OpenVPN they won't.

> a cellular phone; recording even encrypted conversations may be illegal for 
> individuals, but I would expect at least any government would find out some 
> way to justify it. Certainly a foreign government could probably record 
> cell phone calls in the USA, decrypt them, and then pass the data back to 
> the US under various agreements.

You're describing an existing arrangement.
 
> So, while encryption may be useful in delaying people's access to your 
> communications, don't assume it does any more than *delay* access.

It depends on your threat model. For most cases you're wrong.
 
> And it also won't delay access under certain legal situations in some 
> countries:

Now you know why I no longer travel to UK.

> "UK jails schizophrenic for refusal to decrypt files: Terror squad arrest 
> over model rocket"
> http://www.theregister.co.uk/2009/11/24/ripa_jfl/
> """
> So, I suggest we should assume everything we do is monitored and everyone 
> on this p2p list works for the CIA. Now what? :-) Well, one can start 

That is not a very useful assumption. A much better assumption that
everything in cleartext passes through tapping points which can do
nontrivial filtering, and can destil down interesting traffic for further
scrutiny and possible storage, including human analyst eye time.

Already trivial encryption will very effectively disrupt this
large scale passive scrutiny of cleartext. Active MITM takes a lot
more sophistication and can be detected in principle, so it can't
be done on a wide scale. 

> working to change the nature of society in a peaceful, non-violent, and 
> legal way that works for everyone using P2P and other technologies. :-) And 

If you think it will be peaceful, legal and non-violent you're bullshitting
yourself. The current developments are not sustainable. Either the current
trend will reverse, or we'll see considerable mayhem ahead. (We already know
that the current national administrations are deploying this technology to 
early identify and isolate trouble hot spots -- any trouble hot spots).

> one could also work peacefully and legally to changing the laws about 
> copyrights and so on.
> 
> See also:
> "Four Horsemen of the Infocalypse"
> http://en.wikipedia.org/wiki/Four_Horsemen_of_the_Infocalypse
> 
> But we seem way past that infocalypse already in some ways, like the guy 
> above arrested for a model rocket still in the package. In a way, the 
> "think of the children" Infocalypse (like is putting that 22 year old in 
> prison for years) has already come and gone, and we are living in the world 
> on the other side (everything logged, broad and vague laws on all sorts of 
> things that taken together that make *everyone* criminals, selective 
> enforcement of that, children being imprisoned for their own "protection" 
> for sending pictures of themselves, a generated climate of fear and 
> mistrust, etc.).

Precisely. So what are you doing about it?
 
> Still, even assuming the worst, we can work towards something better with 
> what we have, our imaginations, our community, our technology, and so on. 
> Even within the USSR and repressive government, it transformed eventually. 

If you want to assume the worst, you should consider us becoming the
Emergents, in Vernor Vinge's A Deepness In The Sky novel. 

I do not see how we can recover from that once we're there. This isn't
your grandfather's Evil Empire. 

> Things like humor were part of that transformation. So was optimism (one 
> can be paranoid as well as humorous & optimistic at the same time. :-)

[...]

> So, welcome to my paranoid world. :-) Google *is* the government already. 
> Encryption only at best delays access to anything you do online by a few 
> years and at worse makes people investigate you immediately. Everyone is a 

Oh, the horror. The horror of it. Imagine if this was the rule. Imagine
the number of data you would have collect, and how many human analysts
you would tie up listening to pillow talk or academic treatises. If
you've read your Bamford, you'll know that this is not a theoretical scenario.

> criminal because of vague and broad laws that are selectively enforced. Any 
> obvious defiance will get you put away for twenty years in a place without 
> sunlight that may destroy your health and sanity (and that's if you are 

If you are in that situation, it is your civic and human duty to organize
underground resistance. If you don't, you will eventually find yourself at
the receiving end of the system.

> lucky and not just taken out in an "accident"). Now what? How do we build a 
> better world within those constraints? Perhaps by using humor, optimism, 
> and imagination? :-) As well as free-to-the-user Google services? :-)

Paranoia works best if titrated to functional levels. And paranoia and
naivete don't really mix well. 
 
> I'm really a tremendously conformist person in most ways -- you don't get a 
> stamp on your forehead that says "Princeton" unless you are willing to jump 
> through endless hoops on command. :-( My idea of a happy work life at this 
> point is just going to a research lab everyday (an IBM Research like in the 
> 1970s when people like Gerry O'Neill came to give talks about space 
> settlemest?) doing something worthwhile and having a nice meal for lunch 

Not many people are that lucky, these days.

> talking to colleagues. I would not have spent decades outside the system 
> trying to make changes if I did not think they were essential to our very 
> survival (especially because of the risk of nuclear war). Frankly, I 
> probably would have preferred to be a "loser" earning US$200K with the rest 
> of my class (even Michelle); see this humorous item to understand that 
> comment:

[...]

> A "technological fix" with encryption like you suggest does not really 
> address the core issues of a changing society. The core problem IMHO is 

If the society does not use the tools for privacy, the society might find
itself in a failure mode quite difficult to impossible to recover from.

"We have no privacy anyway, so we can give up already" is exactly the 
unconstructive attitude to take. History tends to take a dim view on quislings.

> scarcity-based assumptions at the core of the current political process (as 
> well as economic processes like around even Google), but politicians and 
> business people increasing having to make decisions about using and 
> improving an infrastructure that can produce vast abundance.

Right now the current political and business process is directed towards
surveillance and control of uprisings in the coming scarcity scenarios.
 
> Consider, as one example:
>   http://en.wikipedia.org/wiki/Pirate_Party

I am an active member of the Pirate Party and so should you.

> More on jury nullification:
>     http://en.wikipedia.org/wiki/Jury_nullification

Jury is specific to anglosaxon law, IIRC.

> Guns are like encryption in a sense. Like locks, maybe they help keep 
> honest people honest, but they are not going to do much against organized 
> efforts by government to do something to you. I've tried a few times with 

I happen to disagree about that too.

> conservatives to take them step by step through the idea of how guns may 
> protect their rights realistically in our society and have never gotten a 
> good answer; history in the USA shows the now-despicable internment of 
> Japanese-Americans went relatively smoothly even with a lot of guns around.

The difference between US and EU is that EU does not even have guns to take
away first.
 
> One problem is that the fifth box (ammo box -- which I question the 
> effectiveness and morality of, especially just compared to non-cooperation)

Who's interested in morality when you have a civil war in your hands?

> "Social Movements and Strategic Nonviolence"
>   http://sociology.ucsc.edu/whorulesamerica/change/science_nonviolence.html
> seems to have made many people complacent about all the others. Like the 
> encrypted box, the ammo box does not make anything better by itself. Iraq 
> showed how easy it is to tear apart a society in a way that may take 
> decades to heal. You can hide all the transactions you want behind 
> encryption for a time until society falls down around you, too.

Some severe cognitive dissonance here.
 
> So, rather than focus on encryption, one might focus on making more free 
> content of hight quality, or more free (or even non-free, sadly, like 
> Konarka) designs for technology that produces abundance. Like guns, 

What if this is against the law and will land you in jail?
What if perpetuation of scarcity is exactly the appropriate
control paradigm, and you're the square peg that is in the way?

> encryption may be useful if you want to keep some information away from 
> thieves (who won't try too hard to grab any one person's credit card 
> numbers), but it won't protect anything in the long term from a determined 
> government. Following Manuel de Landa's idea, I suggest we always will have 
> some sort of one, as we have a healthy balance of meshworks and 
> hierarchies. You are trying to deny hierarchies some power with encryption, 
> but they can still get it in other ways. The bigger issue is how do we keep 
> the hierarchies accountable for the health of society?
> 
> So, the deeper issue is, what kind of government do we want? And what kind 

I prefer technology forms leading to empowerment of individuals, not more 
control for superpersonal organisation forms. Encryption as a means for
privacy, strong authentication, trust and tamper-hardening are key parts 
of that of that technology box.

> of society to go with it? And then how do we act to build that society 
> within the possibilities that seem reasonable?

-- 
Eugen* Leitl <a href="http://leitl.org">leitl</a> http://leitl.org
______________________________________________________________
ICBM: 48.07100, 11.36820 http://www.ativel.com http://postbiota.org
8B29F6BE: 099D 78BA 2FD3 B014 B08A  7779 75B0 2443 8B29 F6BE