Schneier on Security

Clive Robinson • December 23, 2018 4:24 PM

@ Faustus,

After a career in consulting I came to conclude that most middle management serves no purpose, as well as most it consulting…

With regards consultants, there are some good there are some bad and then there is the evil that the big accounting firms use to push their methodologies, not solutions to the customers problems.

There is an old truism that “Wise Consultants learn more from their clients than the clients learn from them” and I’ve seen a good bit of that in my time. There is another truism that “A good Consultant only sells the client what the senior manager wants, not what the company needs” which if you think about it is the way a consultant gets invited back. Especially if “A good Consultant only fixes at most eighty percent of the problem, so they can come back and sell another eighty percent endlessly”…

Back a long time ago I made the mistake of “fixing problems” and of “offering to support them” without puting a time limit on it which is why I’m still supporting one or two customers twenty years on. Hence my embargo on “64bit” support with old solutions.

The scary thing however is the Apple ][ stuff was in 6502 assembler and the bulk of it still runs on weird SoC modems and the likes… Much as I would like otherwise the 6502 keeps appearing like “the ghost of Christmas past”… Likewise the Pascal code for energy managment still keeps poping up… Both realy tax the old grey cells when they do.

But “middle managers” well there are three reasons they exist,

1, Emergencies.
2, Organisational memory.
3, Makework / Reward.

Back in the 1990’s there was a self destructive fad called “Business Process Reengineering” Mike Hammer started it but it was Pete Drucker that made it a managment fad.

The underlying argument was thst middle managment represented 50% or more of the wages and other human related running costs in an organisation and that as they only served as information conduits to the C-Suits they could be painlessly removed for greater profitability… Certain major accounting firms with highly lucrative consulting arms pushed the idea as hard as they could down everybodies throat, and with some of their clients it was not optional for them…

What the managment consultants blindly did not see or chose not to see in their rush for fools gold is that middle managment act not just as information conduits but provide translation, filtering, compression and importantly prioritization of the information to basically tell the C-Suits “What they should do and how” in the ordinary stable times. Put simply C-Suits speak an entirely different language to those who actually do the work within an organisation and imported C-Suits are to lazy / full of it to learn to speak to the shopfloor, where as owners and those with serious skin in the game tend to be way more proactive in this area.

It was quickly found that the computers that replaced middle managment were effectively usless due to defective reasoning which was why the 90s were consultant boom time endlessly sorting out (not) BPR falicies and failings untill Y2K became more profitable.

But during BPR companies burned brightly and crashed and burned in a funeral pyre of the consultants making. For a couple of reasons, firstly the smart C-Suits jumped ship to other companies who had not BPRed having worked out what was going to happen next and the majority dumb C-Suits found out the hard way…

Even if there were no emergancies the BPR savings had made artificial profits that speculators loved, the reality was it could not be done safely so to make the same figures again they kept repeating the process and the businesses became more and more fragile untill the first tiny problem caused them to shatter. Because “Dumb is, as dumb always is, and always does, it never learns” these days we talk of the Dunning-Kruger effect or the Peter principle or more modern “Warm Body Syndrome” depending on what the excuse is in C-Row/corridor.

What was happening was that middle managment held the “Business knowledge” which was rarely on the shop floor or up in the C-Row. Once a manager was “outsorced as a unit of work resource” the knowledge and much of the expertise went out the door. But could name it’s price comming back as a consultant… Which quite a few of the smarter middle managers did whilst others setup in competition with their previous employers.

More recently Managment Consultants have been looking at “OODA loops” as the latest “new game in town”. If you look in there you will find that one of the more important things in the “orient” step is knowing yourself. With three of the five points being “Genetic History”, “Cultural traditions”, and “Previous experience”. Which in most businesses is held in the middle managment layer.

But one of the reasons middle managers exist is also about what the OODA is supposed to do which is “turn in tighter circles than the foe”. That is in emergancies middle managers short circuit the information conduit and often act before C-Suits are even aware there is an emergancy. In effect they make the “smaller, leaner, faster” actions that get inside the foes turning circle.

But in times past befor the ludicrous bonus culture kicked in staff that had consistantly performed to the organisations benifit would get actual promotion or new job titles and pay grades whilst still effectively doing the same job. This arangment was benificial to both the employee and the organisation but you had to know why it was there. There were also “Pension Promotions” where final salary scheme workers got promoted to reward long service to the organisation.

Whilst the smarter C-Suite bodies who had significant skin in the game saw the advantages of this system, brought in C-Suite workers want to grab it all now because they have little or no intention of staying with the organisation.

At many levels BPR and many of the systems that followed it worked on the “selfish entity” model not the “Cooperatve Organism” model with the assumption of “Instant hit” rewards and ultra short term thinking. The result being extream fragility in organisations and being easy prey for those who think in slightly longer terms. Which is just one reason the US has given it all away to the likes of China and other BRIC Countries…

Clive Robinson • December 23, 2018 5:05 PM

@ Ismar,

Is it possible to tell encrypted text from the one that is just some random characters put together as one of the security ‘experts’ has recently stated?

The problem is that plaintext statistics do make it into the ciphertext, if they did not then no information could be communicated.

Thus the question becomes one of how much leakage of statistics happens, to which the answer is “it depends” and is one of the reasons for the proliferation of “cipher modes”.

You will often hear said that the One Time Pad has perfect security theoreticaly, well it does not always work out that way when you use a “True Random Bit Generator” (TRBG) to make the pad…

The reason is a TRBG has an “unbounded run length” which means that it could push out twenty or thirty zero bits one after each other, likewise similar length repeating patterns. Which means that in a way you are “reusing” small segments of the OTP which is a “no no” as it alows the underlying plaintext statistics to come through.

There are two basic ways to deal with the problem, firstly clip the excess run length of the TRNG, which might appear counter intuative but works out. The second method is to reduce the statistics of the plaintext. In practice where you can you do both and a little more these days.

So you would do the following,

1, Compress the plaintext.
2, Encrypt with AES in an appropriate mode.
3, Superencrypt with the OTP.

The purpose of the compression is not to reduce the message length (though that may be of use) but to flatten the plaintext statistics. The use of AES in an appropriate mode further reduces other statistics. Thus making what goes into the OTP already to difficult to analyse.

If not is it even possible to tell which encryption was used if you don’t know the plain text?

Due to the way data is transmitted it’s quite often easy to realise which crypto algorithm is in use. For instance with block ciphers the size of the blocks can usually be determined hence give a significant clue as to the algorithm.

Back a century ago when algorithms were ment for the human head and hand to use, it was often easier to tell what language the original plaintext was in just by examibg the ciphertext than it was to determin the actuall algorithm that was used. Often this was because the ciphers were permutations followed by weak substitutions. With the beginings of the more complex cipher machines post WWII repeated layers of permutation and substitution became possible which made the guessing harder but by no means impossible.

Clive Robinson • December 25, 2018 9:55 AM

Trump wants the clocks off

In what some describe as an act of lunacy, US President Donald Trump has decided the Radio Clocks transmitted since WWII should be tirned off.

What he and others around him do not realise is just how many dependent systems there are on these radio clocks. And before people say “GPS” sorry guys it won’t cut the mustard for a whole bunch of reasons.

https://www.voanews.com/a/time-may-be-running-out-for-millions-of-clocks/4554376.html

Not exactly the news you want to hear especially if you know you have dependencies on these clocks, it’s going to be difficult and probably expensive just like Y2K was to sort out. But for others who are not aware if it happens it will be like Y2K happened for them…

Clive Robinson • December 25, 2018 11:53 PM

@ Wael, Tatütata,

Now, 30+ years later, I find out it’s German!

It was housed in a wooden tray, and the springs and devices were retained on a ~2mm thick piece of cardboard.

Yes I had one too, though mine was housed in a cardboard box and more than fourty years ago, it had the polycap variable cap and MW wound ferrite rod antenna, but only did about thirty experiments. It developed a problem that I hope your much later model does not have. The springs were made with very poor quality steel that had poor chrome coating that flaked off leaving a higher impeadence steel that started to rust and become real high impeadence and “noisy as two feral cats in a bag”. Worse was the holder for the AA batteries…

The transistor in mine was if I remember correctly an OC71 equivalent in a black tube that was actually not plastic but black paint dipped glass tube with the transistor “frit sealed” inside. If you scraped the paint off you got an optical transistor that would have cost you three times the price… I don’t actually remember the transistor number but it was a Japanese 2S something. I’d looked it up for it’s equivalent in TITS[1] when pulling the kit appart for bits after the springs had had it.

The annoying thing was the kit was most of the thirty experiments were naff because they had over economised on parts. Whilst it could be either a Medium Wave (MW) Radio receiver or a, medium wave oscillator with “plate modulation” via the little audio transformer it could not be both. Which ment you either needed two kits or a transistor radio.

That transformer you found in nearly all Japanese three transistor MW pocket radios of the time (VHF FM was still not happening back then though it quickly changed). That basically used the same front end radio circuit as the kit but a complementary pair of transistors to give a class AB amplifer at 3V swing to drive the transformer that drove the ~5cm 8ohm speaker.

The kit I had, had no such luxury though, it only had a “Crystal Earpiece” which was actually an old fashioned pink plastic coated “hearing aid” moving coil 70ohm device with pink twisted wire lead. That when you removed the clear plastic “ear tube” became the equivalent of a “moving coil microphone”.

Which right there was a “Security Lesson” in it’s own right about the reversability of transducers…

Years later I used the transformer with a couple of 741 Op-amps to demonstrate you could make a two-wire to four-wire converter that when you put a speaker on the two-wire side, you could both play music through the speaker whilst picking up people talking close to it. It was somewhat of a scary eye opener for some. Back then most of the PI/PS types were advising people that it was OK to talk in a room and use a radio to stop bugs working… It was only in the 1990’s that it became publically known that the Russian cars that carried business men and minor diplomates to their hotels and the radios in their hotels all did this trick. What I’ve not been able to confirm is if it was Russian inventor Leon Theremin who came up with the design or not. It’s known he came up with the design of “The Great Seal Bug” or “The Thing”[2] as it was called as well as that weird electric musical instrument that carries his name and can be heard on the Beach Boys record “Good vibrations”.

As for those transformers they are like “gold dust” these days, I needed one for a little audio issolation box the other day to hook up my son’s Yasseu all band QRP rig (FT-817) to his “old” smart pad so he could run some digital modes “portable”. Hunted high and low even for old PC Modem cards in the scrap box to get one from but I’d used the lot. Ended up having to use a pair of back to back electrolytic capacitors[3] and a current biased BC109C as a low gain Class A buffer amp. I hate current bias as it’s prone to not just thermal run away but becoming an oscillator[4]. Worse still for portable work it needs a power source unlike a transformer 🙁

[1] Towlers International Transistor Selector or TITS for short was a yellow paperback “pocket” book about A6 in size. About 2cm thick it listed equivalents for the 2S Japanese 2N US and Mullard BC etc transisters so repair technicians could find replacments. Though experience taught me that BC107 through 109 and later ZTX300 usually worked in voltage biased circuits. Current biased was as they say “A,whole different kettle of fish”[4]. Whilst I still have TITS kicking around the workshop somewhere I no longer have the compleate Mullard Valve data books that had glossy A5 sheets of paper in hardish cover black leaver arch ring binders. They covered every octal base valve (tube) they made, and some realy weird bottles like the Line Output Devices used in colour TV’s that had X-Ray warnings (they actually glowed faintly “apple green”). The pages and the binders they came in had a characteristic smell, I later found out was due to a nitro cellulose finish that was degrading into something that would become a mixture of a nitro based explosive, and fire accelerant…

[2] I still design surveillance devices that work on this principle though they are quite a bit more advanced than the version you can see in the TAO catalogue. The original or a duplicate is on display in the NSA museum,

https://www.cryptomuseum.com/covert/bugs/thing/index.htm

However what the crypto museum says about the UK’s Diplomatic Wireless Service (DWS) and the UK’s MI5 involvment in the incident is unsurprisingly “down played” read Peter Wright’s “Spy Catcher” for a lot more of the story, the web site is being overly shy about. Sadly Tony Sale (of Bletchly Park fame) is nolonger with us. He was Peter Wright’s assistant and could confirm the part MI5 played in things. The website talks about a “Crystal video-Receiver” this was an 1N23 diode detector followed by a wide band “video amplifier” that eventually got upgraded in the US by the addition of a low power Traveling Wave Tube RF amplifier. Long after the DWS had moved to rather more sensitive equipment. It might well account for why the DWS first detected the Great Seal Bug output signals and reported then back to the US, through a route that avoided the majority of the US IC and SigInt entities (some things are just to important to tell your friends who gossip).

[3] The use of back to back electrolytic capacitors to get over the polarisation issue is something that should be discoraged in designs these days. Because not only are modern electrolytics way less tolerant of this abuse it causes increased noise into a very high impedence to get a flatish frequency response, thus is not a good idea with low level signals. Yes you can use one polarised electrolytic capacitor, but only when you know which way the DC offset is going to be, which when you are going for full “galvanic issolation” for electrical safety reasons does not happen…

[4] Proving the old design engineers saw of “Oscillatots don’t but amplifiers do…”

Clive Robinson • December 26, 2018 6:27 AM

@ Wael,

My goodness, you’re gonna get me banned!

Worry you not it’s a genuine book in goodness knows how many editions in it’s various sizes.

It would appear you can buy one for New Year and have Amazon drop it in on you by low flying death bot or whatever their drones will be called 😉

https://www.amazon.co.uk/Towers-International-Transistor-Selector-Tower/dp/8170297982

Clive Robinson • December 27, 2018 5:07 AM

@ 65535,

So, the DOS is on the set in separate disks?

Yes even in some early versions of Win95 you could easily pull out the MS-DOS Component.

Few realise that having “Windows as an overlay” was happening with Win-NT. That likewise is very similar in that you don’t need to run the Windows component, and back in NT4 many people ended up doing just that on servers due to the “Screen Saver bug”. Put simply your WinNT server was doing greate and you were happy. You’ld then go get a fresh coffee sit down to contemplate what you were going to do, when the phone would ring and either a user or first line tech support –depending on the size of the organisation– would tell you your server was down or slower than a geriatric snail on mogadon.

You’ld hastily reach for the mouse and type in the admin password and pull up whatever instrumentation software you were using…

It did not take long for people to realise that the screen saver was swallowing upto 90% of the CPU performance…

The oft used work around was “Disable the screen saver” the down side of this was the security risk…

Speaking of which look out for MIcrosofts very own version of the *nix chron-shroot bug oh atleast a decade after *nix had solved it’s version.

Put simply you would login as an ordinary user set up a job to restart the win shell in oh three minutes then kill your current unprivileged win shell to get back to the dos shell without loging out. The job runs the new win shell starts and you are in it, you check and now find you are “system” as the user…

This was just over a decade ago and has been subsequently fixed. But it did start me thinking about “Why are we not learning from past solved problems” which unfortunately is still happening as I comment from time to time 😉

Clive Robinson • December 27, 2018 11:41 PM

@ Rach El,

I keep hoping to hear my favourite phrase, Einstürzende Neubauten

Why imploding new buildings?

Or do you just like the way it “rolls off the tongue”?

Clive Robinson • December 31, 2018 10:17 AM

@ ,

That portable you have locked up that you recently mentioned wouldn’t happen to be made by somebody named Alan would it?

Yes if you mean “Your Fired!” man.

For those that don’t know the popular program “The Apprentice” started in the UK and the boss used to have a company called Amstrad, since then he has moved into “property” and has picked up a “Lordship” along the way.

He also has some rather nasty political views these days that can make the UK far right look almost like fluffy bunny huggers in comparison…