Slashdot is powered by your submissions, so send in your scoop


Forgot your password?
Canada Government Security Your Rights Online

Thank Goodness For the NSA — A Fable 60

davecb writes "Slaw was kind enough to post my fable on how to not have a problem with the NSA, Thank Goodness for the NSA, and a link to the more technical MAC paper. My challenge to the Slashdot community: what's the first big step to making this all come true?"
This discussion has been archived. No new comments can be posted.

Thank Goodness For the NSA — A Fable

Comments Filter:
  • I'm just going to assume it's the following things in equal measure:

    A. Poorly written satire, where "thank goodness for the NSA" is a repeated statement made by forest creatures, ironically unaware of their own doom they weave.
    B. The positions of the satirical critics are 100% holy and just, but no one believes them.
    C. Believe in the NSA, not apathy is the driving force of its existence in the story. And..
    D. The moral reinforces my biases.

    • Re:Blocked at work (Score:4, Informative)

      by Anonymous Coward on Thursday January 02, 2014 @12:08PM (#45846321)

      The Article:
      If it weren't for the U.S. National Security Agency's trying to spy on everyone in the world, Bleeker Street Law would have been a cooked goose.

      Back in 2013, we had a group of clients from a particular country applying for refugee status here in Canada. Because the NSA spying was in the news, we did a forensic audit of our computers, just to be safe. We promptly discovered that we had been hacked. Not by our clients' former national security service, or by the NSA, but by a for-profit organization. A set of aspiring criminals had broken our security and were making everything they stole available by subscription on Silk Road. Several foreign firms and at least one government had subscribed to us. . . .

      The country in question had a revolution, Silk Road doesn't exist any more, and we now have a much simpler but more secure computer system, mostly on tablets and phones.

      What we do differently

      We used to worry about privileged communications with our clients, because we did all too much communicating with ordinary unencrypted email. Now we have encryption programs for our pads and phones, and encrypted email to boot. Older machines storing files get them already encrypted, so crooks can't just subscribe to “every updated file”.

      One new machine keeps the keys. We guard it like the cabinet of office keys, and it in turn is locked in the law librarian's office and not connected to networks.

      What's on the pads?

      Pads are very popular, and both Apple and Android have “end to end” encryption programs on them. This allows us to “label” files with encryption keys, so only the right people can decrypt them.

      Personal information is labeled with the person's name, which in effect means it is encrypted with the person's personal key. Business information is labeled with both Bleeker Street's name and the name of the person whose pad or phone it is on. It is therefore encrypted with a per-person business key.

      Only little bits of data are in memory and unencrypted at any time, and because it's labeled, it's re-encrypted when it's written back to disk..

      Clients can download a free app and have secure email labeled “From client, for Bleeker Street”. We have the for-pay version and can talk to them and to each other, using keys that live in the locked machine.

      What's in the keystore?

      Our keys, starting with a private key for each of us, then a collection of public keys from our staff and clients, and finally a collection of keys, each of which is for the combination of Bleeker Street and an individual staff member or client. We also have some signatures for software we use (we have a secure subscription), certificates for web pages and the like.

      A legitimate investigator can get a court order to get individual keys, but they won't get all the keys and therefore individual lawyers and clients aren't at risk from them.

      Where's the risk now?

      Stealing data while it's in use is the big risk, followed by people shoulder-surfing for passwords when they're typed. The labeling of accounts keeps most data safe from anyone other than its owner, but if someone subverts the machine itself, they can get data from memory and tiptoe away with it.

      It's not perfect security, but we're not an attractive nuisance any more. Criminals used to target us because we had lots of valuable information in one place. No longer: now they have to attack individuals.

      They still do, mind you: someone tried to claim they were a partner's daughter in a foreign jail last week; but they can't just break into a file server and take the company's crown jewels. If they do that now, all they'll get is encrypted files, which are about as valuable as zircons.

    • Re:Blocked at work (Score:5, Insightful)

      by crashcy ( 2839507 ) on Thursday January 02, 2014 @12:09PM (#45846329)
      E. Nothing to do with the NSA, and not a fable. His company's security sucked, they got hacked, the improved their security. That's TFA.
      • Well, yay for corporate censorship combined with misleading headlines, then.

      • by Anonymous Coward
        Having worked with some law offices (lawyers individually too), and seen their complete blase attitude to the information that they hold in their files, this surprises me not at all.

        Too many attorneys think that because the law says their information is private that it is so, and are absolutely shocked when the other side produces confidential conversations that went through gmail or some other source.

        Truly astounding.
      • Re:Blocked at work (Score:4, Informative)

        by wvmarle ( 1070040 ) on Thursday January 02, 2014 @12:48PM (#45846717)

        Except for this little bit in the italics below the main article text:

        All of the capabilities mentioned are real as of 2013, and have some degree of availability. No-one has a product that provides them all as yet. Full disclosure: I once proposed this to a device manufacturer, who thought no-one would ever need it.

        Sorry, it still is a fable.

      • E. Nothing to do with the NSA, and not a fable. His company's security sucked, they got hacked, the improved their security. That's TFA.

        Actually, it does derive directly from the NSA. Specifically, it comes from the NSA's research on Mandatory Access Control, which is the theory underlying all that discussion of "labels". MAC doesn't necessarily use encryption; in its original design it was intended that the operating system enforce the access controls, but it actually matches quite neatly with the capabilities of labels which correspond to private keys.

        So the fable (I agree that it's not a fable) is about using NSA-developed ideas to sec

        • by dryeo ( 100693 )

          This is a Canadian law firm. The NSA is supposed to be weakening their security so they have data to trade with Canada's 4 letter agencies for data on Americans. This way everyone can legally spy without breaking those pesky Constitution things.

      • by davecb ( 6526 )
        Alas, it is a fable: the story is set some years into the future, when such capabilities can be bought off the shelf.
    • by mlts ( 1038732 )

      It mainly is an emphasis on going with endpoint security... something which should have been done well before the NSA came to be a boogeyman.

      Of course, the article glosses over the biggest gotcha of endpoint encryption... key management. eDiscovery is a major part of business these days, and having a way to recover documents is often mandated by some regulation.

      For a small law firm, this isn't a big deal. You get all employees to send stuff out that has the firm's ADK (additional decryption key) attached,

  • by dmbasso ( 1052166 ) on Thursday January 02, 2014 @12:04PM (#45846281)

    The actual title should be "thank goodness $SECURITY_THREAT made use realize our security was worse than crap".

    • by return 42 ( 459012 ) on Thursday January 02, 2014 @12:17PM (#45846403)

      I think it's more like, thank goodness $POWERFUL_PERVASIVE_SECURITY_THREAT made everyone realize their security was worse than crap, because otherwise they would never have gone to all the trouble of fixing it. Plus various suggestions for how to fix this state of affairs.

      He seems unaware of the issues with compromised hardware, which will require either a political solution or a whole lot more work than software solutions, but as a call to action, it has some merit.

    • The actual title should be "thank goodness $SECURITY_THREAT made use realize our security was worse than crap".

      Well if they're using Perl in their security software no wonder they got hacked...

  • by ka9dgx ( 72702 ) on Thursday January 02, 2014 @12:13PM (#45846357) Homepage Journal

    So, what these articles are both calling for is Capability Based Security, in which you feed a list of resources to the OS when you run a program. This has the pleasant and reasonable effect of limiting the side effects a program can do, and protects the user, the operating system, and everyone else on the internet.

    The trusted systems of the 1980s required the Administrator to supply these lists... it could reasonably be done by users these days, because we're all system administrators of our own machines, when it comes down to brass tacks. It doesn't even have to look much different than what we're used to seeing. A capability based version of Word would ask the system to get a file... which would do so via a "powerbox" (a secure way of picking files which side-steps the application doing it directly).

    I applaud this fellow traveler who seeks the same sane approach I've been shouting about for years. 8)

    • More over, it doesn't even need to be something complex as a per register, memory block level. It can be as simple as saying "this is a game" and "most games don't need root file access" therefor "this thing that calls itself a game, is trying to use root file access, stop it."


      Most graphics tools are the same in what they touch. Most games are. Most [insert anything].

      It wouldn't be hard to implement, or specify. Hell, many applications are ALREADY categorized. You press "apps" button or the
    • by chill ( 34294 )

      SELinux (and SEAndroid)

      ...because we're all system administrators of our own machines, when it comes down to brass tacks.

      No, you're not. If you allow that it is Discretionary Access Control. There has to be a master policy that is enforced, limiting DAC to only where appropriate.

      For example, if Alice allows access to Bob and Charlie, she should have the ability to restrict resharing. That is, unless Alice allows it, neither Bob nor Charlie could grant access to Dan. Therefore neither Bob nor Charlie is a true admin of their machine.

      • by naasking ( 94116 )

        For example, if Alice allows access to Bob and Charlie, she should have the ability to restrict resharing.

        Why? This access control merely provides a false sense of security. Bob and Charlie could always just make a copy and send that, or proxy access for whomever they wish to share to. It doesn't actually restrict delegation, it merely provides the illusion of doing so.

        Restricting delegation has caused all sorts of insecure mechanisms just re-enable the types of sharing that are needed just to get work done

        • by chill ( 34294 )

          Copying and providing proxy access are process controls. You discipline people for that.

          You're demanding something perfect and rejecting anything that doesn't measure up. The real world doesn't work that way. Security layers are added depending on the specific needs for confidentiality, integrity and availability.

          Something like the formula for Coke or the KFC Original Recipe get more protection than the cafeteria's lunch menu.

          • by naasking ( 94116 )

            Copying and providing proxy access are process controls. You discipline people for that.

            No, it's worse than that. Firstly, they are efficiency impediments because they require workarounds, and moreover, they obstruct the deployment of fine-grained permissions ala POLA because users still need to be able to do something to do their jobs. Secondly, training users has never worked and will never work, particularly when such discipline conflicts with the need to do their job.

            There's a reason capability-based se

    • by PPH ( 736903 )

      Yeah, yeah.

      because we're all system administrators of our own machines,

      And by 'we', you mean the majority of users who see the UAC popup and only understand, "Blah, blah. Blah, blah download, blah, blah, blah. Yadda, yadda, codec, blah, blah, blah." And all they want is a button that says, "Make the nasty box go away and give me the cute kitten desktop."

      Game over.

      • by ka9dgx ( 72702 )

        Not so fast... we all administer our own wallets, and we know not to send all our money to a PO Box in Nigeria. If there's no way to specify what can/can't be accessed, you get the default behavior you describe, because there really is no control. What we have now are systems just like EULAs, you either choose to run a program, or don't.

        It doesn't have to be this way, and it wouldn't even cost much more to do it right. We could all have Orange Book A1 Secure computers, if we wanted to do the work as a com

        • by PPH ( 736903 )

          But most people understand PO boxes and know where Nigeria is. Grandma just thought the kittens were cute and had no understanding why a binary from .ru would be something to worry about. OTOH, grandma has come close to cleaning out her savings account because Leisure Suit Larry said he'd hold the cash for her.

    • by davecb ( 6526 )
      To be somewhat nitpicky, capabilities as discussed here are a way of enforcing categories. They're main value is what you mentioned, their ability to put fine-grained restrictions on processes, such as "you're a game, not a debugger, so you can't read another processes memory".
  • by Chalnoth ( 1334923 ) on Thursday January 02, 2014 @12:13PM (#45846367)
    Thank the person that brought these security breaches to light, not the people who have been illegally performing them.
    • Hear, hear. If I had points you'd get one.

    • Edward Snowden had jack-didly to do with bringing any security breaches to light. If a company like this one didn't do regular security audits, then that's why they got breached in the first place.
    • by Anonymous Coward

      I think the point of thanking the NSA is that they were the root cause - Snowden was the proximate cause. The issues for the author of the original article did not stem from the NSA activity, but the NSA activity led to Snowden's revelations which led to the author's organization examing their security and discovering that they had been hacked (by organizations other than the NSA or Snowden). If the NSA's activities hadn't inspired Snowden to perform his expose, then the author's organization would have g

      • by Anonymous Coward
        No, the point of putting NSA in the headline and article (twice) was to get hits on the /. editor's keyword filter. Anything with those three letters will get auto-greenlighted until it blows over. Relevance to the actual story or to /. is irrelevant.
    • by davecb ( 6526 )

      Strongly agree!

      [Thanking the NSA was just a tiny bit tongue-in-cheek]

  • by BringsApples ( 3418089 ) on Thursday January 02, 2014 @12:39PM (#45846621)
    Try to convince yourself that you didn't just get tricked into reading an article.
  • If it weren't for our own inept network security implementation, all are eggs would have been in one basket.

  • Hitler, the Black Death, Attila the Hun, Toba, the Chicxulub asteroid, whatever caused the Great Dying and so on. That we survived despite (at a very high cost) them don't mean that we must be grateful for what they did, even if that meant that had a role on the changes that ended with us right how we are now.
  • The one machine that has all the keys is in a locked office, not connected to the Net.

    Lessee, 1) do they *also* have an offsite backup of that info in a safe deposit box somewhere?
    2) if not, and there's a fire, what happens to their company?
    3) Who installed the lock on the door? Does the building engineer have a key? How does

    • by davecb ( 6526 )
      Yup, in the Fabulously Secure Supplier Company's safe (;-))
    • You assume the system works. In the real world, the keys cannot be just in a computer that is not connected to the network, because otherwise nothing would get decrypted outside of said computer.

      The moment you try to get those keys to be useful, you are making copies, which will reside in less secure environments, and you are connecting something to said machine, which suddenly gives you an attack vector.

      A system that has very secure keys, like something that meets PCI-DSS key management standards, is alway

The optimum committee has no members. -- Norman Augustine