Schneier on Security

Clive Robinson • November 21, 2015 10:16 AM

TrueCrypt users may be interested in,

http://arstechnica.com/security/2015/11/truecrypt-is-safer-than-previously-reported-detailed-analysis-concludes/

It appears that the 77page report on the audit is now available.

Yes new bugs have been found but the report authors indicate they do not consider them of serious risk (others will no doubt beg to differ).

Clive Robinson • November 21, 2015 6:40 PM

@ Nick P,

The prob with the PISC is four 181’s and eight 172’s, if you try and find some your eyes might water harder than if some one pulled your wallet out of a lower orifice.

I’ve a few but even I’m asking $150 each as they are mil spec and still in hermetic packaging. Just be thankfull they are not in the 29000 range, the mil spec version of those I want real money for as they are worth more than their weight in 999fine. Such is the game of having a “Computer Museum” worth of spares in the loft and other places. Mind you my early IAx86’s lost value when NASA pulled the STS, so it’s swings and roundabouts.

But you probably aware of that. As the article pointed out DIY CPUs are not easy to do these days as the chips you need to breadboard with are rarer than hens teeth. So VHDL and FPGAs are the closest most students get if they are lucky…

However there is another way for those with a little patience and carefull thought. You can use AND-OR or OR-AND matrix chips such as PALs (getting scarce) or RAM/ROM. I’ve had fun using nS speed SRAM chips that you can load with a little PIC chip with USB connection. The cost of playing this way is quite minimal, you can also with a little thought make the same circuit do other interesting tricks such as a fully digital 10.7MHz digital IF line up to feed into a high quality sound card to further process on a PC. Thus you don’t have to spend a fortune on the care and feeding of specialised chips and laying out expensive custom PCBs etc.

Clive Robinson • November 22, 2015 7:31 AM

@ Jacob,

Of those parts listed non are mil spec or gold lead ceramic package. In Europe it’s doubtful you could sell due to PbFree legislation. Also some may be second hand at best (many are pulled from DIL sockets but also some from old PCBs and only given a cursory test despite having a real bad time being desoldered, cleaned and reformed).

All of which makes a very significant difference to the price, and also to the performance in some cases (buyer beware, where there’s a market with lots of profit in it there will always be scammers and crooks)…

The 181 with just under eighty gates does not make it into the MSI category. Thus when you normalise the price per gate you quickly see why they are not used much. Also most “old stock” got sold off in the 80’s and 90’s and were quickly crushed and melted down for their very high precious metal content.

But it gets worse… to do an 8bit data, 16bit address processor you need two 181’s plus several other control “glue logic” and shift chips. That’s all before you add four or more 172’s register files and their glue logic.

Once you have this nest of around 20 TTL packages, then you need to start thinking about the RTL and instruction decode control logic. You get a trade off between RISC and CISC designs and the quantity and speed of cache memory and core memory. Old style went CISC then as cache tech improved RISC. But these days it’s both, with a RISC Harvard core and level 1 caches, then a CISC instruction decode segmentation –poor man’s MMU– and data layer queuing/reordering and level 2 cach all wrapped up in a Harvard to von Neuman bus multiplexor to save on external pin count and pin driver/protection real estate and gain extra flexability (as well as security head aches).

To give you an idea of what was involved, the Z80 8bit CPU was actually a 4bit ALU with some very interesting “dynamic logic tricks” pulled (hence the minimum 0.2MHz clk speed). But even then it would need a hundred or so TTL chips to replicate. The 16bit 68000 design prototype I was shown years ago, was wire wrapped and filled three full size Euro Cards and could barely hit a 1MHz clock speed reliably using AMD 2900 parts. The first 68K chip hit 10MHz performance in the correct packaging.

Importantly though when you normalise the gate price on the glue logic you are up and beyond the 10cent/gate cost… Modern 8bit CPU’s way in at 1500-4500 gate equivalent depending on who’s figures you believe. But the chip price brings the gate cost in at less than 1/40th of a cent. If you could buy the bare CPU, which you can’t. Now you get a lot of free memory and IO which multiples the gate count by several orders of magnitude and thus brings the gate cost down, way way way lower than that, such it’s not used as a measure any longer (remember to knock a dollar of the price for the packaging).

I’ve built boards using modern SRAM chips and PIC micro controlers to load them, that can emulate most of a 68K at a considerably greater performance than it’s possible to build with TTL. Importantly if you make a mistake in your design logic, you just hit reset resynth the memory map in software in a few seconds… and reload.

Which is why I don’t recommend anyone trying a TTL only chip solution for anything other than for bragging rights or to entertain others. Look at it this way, for the cost alone you could get a decent off road bike and have a healthy outdoor hobby that would do more for your body and mind than sitting hunched up, going short sighted and bloodshot, breathing in solder fumes and being in worse health than a two pack a day smoker…

And for those that realy must build a hobby ALU/CPU can I suggest you design a “1Bit static serial ALU” with still readily available and thus cheap parts and use it to teach / entertain with. I’ve built one with relays led displays and a 0.5sec clock which gets used at science / technology shows etc to show what a “school project” can be, and even mobile phone clutching teenagers stare in fascination at it working to add just a couple of numbers…

Clive Robinson • November 22, 2015 7:00 PM

@ Justin,

Tin and cadmium are two metals that have caused many problems in solder over the thousand odd years man has been soldering.

Tin rot has “unknown components” but what is known is that it sufferes from an allotropic change at the bottom end of acceptable non specialised clothing temperatures (14 +-3C). That is tin is known to go from a silvery metalic conductivr form to a dull metalic grey crystal/powdery non conductive form. Which has been specifically known by organ pipe makers since medieval times. As with all amalgams of metals (including the fillings in peoples teeth) there is a great deal we do not know.

When the RoSH rules were originaly discussed the makers of musical instruments said that the removal of lead from soft solders would be a mistake the industry would quickly regret for exactly this reason… History has tales to tell with “uniform buttons” although it is unlikely that Napoleon’s soldiers died due to “tin rot” in their buttons, tin buttons in warehouse storage are known to have been destroyed by “tin rot”. What is not known is why it appears to spread like a disease and little or no research could be the main factor rather than false recording of past events.

The specialised use of tin is also known to be an issue with other electronic components containing purer forms of tin such as LCDs where thin layers of tin are laid on glass so as to be near invisable to the human eye but still be conductive.

We need research done, as the lack of knowledge has cost us big time already. Atleast one very expensive satellite has been known to fail due to the odd effects of tin. Thus caused it to be baned from use, and thus replaced with other metals that can add 100Kg of weight to a standard space payload, which adds a mind boggeling cost to launches.

But there is also the primary issue of “what happened”, it is suspected that a tin wisker caused the computer to fail due to “shorting out”. Tin wiskers have been found in both conputers and communications equipment that our moder society is built on… Other recent evidence indicates similar tin whiskers have caused “sensor failures” in vehicles and thus accidents with fatal outcomes. Do you want to sit in the seat of a car which might have a non mechanical brakes failure at 70Kph? How about an aircraft with auto-pilot or engine/flight control failure? What about having a medical implant such as a pace maker fail? Or smart meter in your home that decides it’s 40 above when it’s actually 40 below and forcefully cuts the power to the heating? So definitely it is a subject that falls into “computer security” just on technical grounds.

But how about political as well?

The existance of tin rot in it’s various forms has been known about for some time, and yes the warnings have been ignored for what some consider are incorrect “political” choices. As is normal in such choices “science” was not involved in the outcome and as a result the wrong choices have probably been made…

Whilst there is nothing new in this history tells us yet again that using knee jerk politics and not considered and reputable science to change societal behaviour is going to cause ever increasing problems with time (such is the way of the world).

Oh and further highlights a problem in politics in general. There is a standard political argument that “The men of science will find soloutions” well they don’t get the chance pute and simple. Because “the men of quick profit buy the legislation” more often than not. But perhaps worse bad legislation happens because we also “think of the children” and plan by worst case scenarios thought up by those with an agender and not science, mathmatics or more importantly reason. That is an “outlier event” causes a news story which in turn causes not just FUD but a sense of agrievement and thus an agenda.

People with an unreasoned agenda are suprisingly quite common but they also tend more towards extreamism, which attracts in others “looking for a cause”. As has been seen in more recent times the Internet alows them to feed on each others needs which attracts in others with psychopathic / sociopathic intent and cause what we now call “radicalisation” and terrorist type activities.

So whilst at first it might appear that “tin rot” has little to do with computer security, it can be seen it’s only one or two degrees of freedom away from it, both technically and politically, and further similar things can with a little FUD thrown in the mix become just one more degree away from violent activism and thus tip over into terrorism. Which has been seen in the past with medical and other research in the west.

This is known to people in authority, there is a standing joke about the UK Home Office. When a new Minister of State is appointed they ask the obvious question about what is likely to be an issue etc. Supposadly the Civil Servant in charge invites them to look out of the window at the weather and predict when it will next be blue skies, grey or even stormy weather, and comment that even the weathermen don’t know from one day to the next.

We even have a saying about a “butterfly’s wings” in some distant place causing storms above our heads. The point being that even the tiniest least likely thing could become the next “cause celeb” for a “nutter with a bomb”. Hence the saying about “Three sign posts to disaster”. The first can only be seen with hindsight, the second only to the worldly wise and acutely observant, the third obvious to all but those of the dimest of perception or intimately involved.

It is this first and second step that “collect it and store it all” is supposed to make easier. Hence the “time machine” effect they are trying to build at vast expense. This comes from the same flawed arguments that can be found in the supposed advantages of “High Frequency Trading”. Which give rise to the appearance that knowing things just miliseconds before your competitors gives you a massive advantage. The problem with this is reality, HFT is a faux market place with it’s own faux rules based on different assumptions than the real world but tainted by the idea they are related in ways they actually are not. Using the ideas outside of the faux environment will more often than not lead to failure or less frequently a world of hurt for the innocent, with success being no more or less than predicted by a roll of the appropriate number of dice. But how many dice? We have no idea if it’s twenty, a hundred, a thousand or million dice… But with just ten giving odds of over one in sixty million we can see that even vast stores of data and computing power is not going to be even remotely of help in getting a timely solution. That is assuming of course that we knew what were the right questions to ask in the first place… Which for obvious reasons we don’t…

Clive Robinson • November 22, 2015 9:18 PM

@ Dirk Praet, Justin,

In the past there has been an argument that terrorist commanders do not want survivors from attacks coming back to haunt them.

That is they are only suitable to be martyred not congratulated. Their mentality is unsuited to becoming teachers, and their minds will dwell on what is in effect their failure to fall in battle and ascend to heaven in a state of grace, which was their personal objective instilled in them during radicalization. Thus they might start to question their teachers and that makes them a danger to their superiors, that can not be allowed…

But there is another asspect, a dead terrorist is often a dead end as far as enquires go. Because little was known about them other than they dropped out of sight some time previously, without ever having made it onto the radar, now they are just bits of a body in a bag destined for an unmarked hole in the ground.

But being a live known criminal and very much target zero on every bodies radar is a significant problem. The sort of support network required for that is orders of magnitude more difficult as there are no safe havens these days for marked men, death and destruction follow their footsteps. There are reasons why drones are known as reapers and it has dark humor behind it, not lost on those who run the terrorist training or those that assist them. These people have lives, families, freinds and importantly a sense of self worth of being sufficiently important to the cause to justify rather more than their continued existance.

The simple fact is those who run terrorist networks these days don’t have the ability to train opratives to the required level of OpSec that they can survive on the radar. Nor do they have the ability to train those who would need to act as deep cover recovery specialists, even big states have real troubles here which is why NOC officers are rare and contractors payed in ways that you or I could only dream of (think the best of “insider trading” and full tax cover).

Getting involved with such terrorist organisations is not a glamorous life or a life at all, it’s a death sentance these days. There is soon going to be no way home for jihadists, they will not be wanted there or at home. They will be stateless and without assets to survive upon.

The reality is their usefulness is non existant there other than as cannon fodder or equivalent of a slave. They are “marked for death or explotation” one way or another the moment they arive and will be shunned, killed imprisoned or effectively enslaved if they try to stay, or imprisoned if they try to return.

That is their journey once started is a termination of their existing life, with no way back and as for forward… like most westerners they don’t have the life skills to survive in a world they can not comprehend or have the resources to buy protection in, and where religious piety will not put food in their belly, but will get them stoned to death if deemed insuficiently pious.

So how they are going to survive where they have no appropriate life skills or welfare or family support, don’t fit in with the culture or speak the language as anything other than an outsider, is something they have obviously not considered. I guess neither have they considered disease, infections or accidents in a part of the world where a minor cut can kill you for want of clean water, a bottle of antiseptic, sterile stiches, a handfull of pills or a health care worker. Where many more young women die in complications of child birth and malnutrition needlessly than die in accidents. Where guns and knives settle arguments and people just disappear with barely a comment with out a family to push and bribe what passes for authorities to get questions asked, let alone answered.

Clive Robinson • November 23, 2015 12:00 AM

@ Nick P,

You may find this of interest,

http://www.utkusen.com/blog/destroying-the-encryption-of-hidden-tear-ransomware.html

Clive Robinson • November 23, 2015 8:56 AM

@ Dirk,

He’s still at large and believed to be in the possession of an explosive belt he may yet use against some soft target in the Brussels area.

Has there been any word on the type of explosive in use?

It’s just that some homemade explosives age very rapidly, becoming either to insensitive and thus not going high order, or to sensitive and going high order if you just look at them hard.

Thus if made with either non comercial explosives or detonators the belt could now be or will be soon of no use to him in attacking a target.

I could be wrong, but just maybe he’s waking up to the “real world” and realised he’s a bit brighter than the mindless programed cannon fodder. And just blowing yourself up as a couple of the Paris attackers did at the football stadium is not the way to go…

Perhaps mad as it sounds he might ask for and get special protection. HE must know that the authorities will kill him if he makes just one mistake, but that they would rather have a long chat. He also probably realises that his former colleagues want him dead so he can not have those chats… Thus he may realise or be told he’s now in the same position “as a silver fob watch at a pickpockets convention”.

Clive Robinson • November 24, 2015 10:58 AM

@ Czerno,

You definitely don’t want to experiment with those compounds in your mum’s kitchen !

Not if you want to keep the roof on the house…

Acetone is a quite volatile fuel, and is not difficult to get hold of, as it’s used as a solvent for many things, including acting as a “cold weld glue” for many plastics.

You also do not want to breath in the fumes as to quote some manufactures “It has known toxicological disadvantages” which is a nice way to say it kills you quickly by inhibition of biological processes such as resperation slightly more slowly by other lower dose inhabitions or at microscopic longterm doses by causing cancer etc…

Hydrogen peroxide is really nasty stuff and during WWII the Germans used it as an oxidiser for rocket fuel… Basicaly it does not play well with anything organic. Whilst low concentrations are not dificult to get hold of, the high concentrations are quite difficult.

Over the years “recipes” to concentrate it have surfaced in things like the “Anarchists Cookbook”. People following them have usually not survived the process unharmed.

Needless to say the result of brewing up this fuel and oxidizer with the appropriate acid and low temprature is one with one heck of a lot of energy stored in it’s chemical bonds. But… Interestingly when it does explode it tends to form acetone and ozone, not oxidation of the fuel. I’m told it was at one time considered for use in Fuel Air explosives by the Russians because of this.

Further if it is not stored correctly it subimates with a loss of around 50% in around 250hours. Also the sublimation is quite an unstable process and can cause spontanious detonation. Which means that although it’s a primary explosive –you could use in detenators and firing caps– it’s not stable enough to be used as any imperfection in sealing could end up in spontaneous detonation as could just leaving it on a shelf…

Which brings us around to what has not been talked about, which is the kind of detonator they used. This is the key component in bombs in the same way a match is to a bonfire or barbeque. Whilst “hot ignition wire” can be made with old fashioned flashlight bulbs it takes skill and practice to get them to be reliable with most primary explosives. That is it’s not just heat you are looking for but the right sort of temprature rise, to fast and the wire will not transfer sufficient energy before it burns out, to slow and you don’t get the primary explosive to give the required pulse to the main charge to ensure a high order result. Acetone peroxide is a little more forgiving in this respect but as previously noted it’s asking for trouble doing so due to sublimation.

At one point the main advantage to acetone peroxide other than the easy access to precursors was it had no nitrogen content, which the majority of CAM devices looked for when detecting explosives. However acetone peroxide smells strongly of “bleach” or “acetone” thus it’s quite easy to detect due to the sublimation or wetting evaporation it requires.

Clive Robinson • November 24, 2015 12:57 PM

@ Nick P,

Something you might enjoy….

http://newsthump.com/2015/11/24/windows-10-gains-self-awareness-initiates-judgement-day/

Clive Robinson • November 24, 2015 5:13 PM

@ Crusader,

That is a guess some have made. The US disbanded the Iraq army. Many were Sunni. Daesh is Sunni.

It’s quite a bit more than a guess. These people have not been in hidding, and on some cases their position has been verified.

The problem I have, is if you are going to send a drone in to kill people both intentionally and with expected collateral damage. Why go after a military insignificance like “jahadi john” when you can go after the known commanders and leaders.

In terms of Return On Investment on a drone strike, offing a military commander who can not be easily replaced and offing a psycho with a sword who is a ten a penny nonentity, the military commander should win hands down, but for some reason –political mileage perhaps– did not.

As for Anonymous, well it does look like a PR stunt more than a serious indent into IS’s operations. But these things rarely happen in a vacuum so there may be some form of blow back down the line.

We have seen this with Iran as a result of Stuxnet. Whoever the Iranian hackers are, they are getting more and more proficient and they have caused real problems (Saudi Oil, erasing of 30,000 computers is a fair accomplishment in any persons book).

I can see the advantage of IS recruiting hackers in western nations to attack localy, a thought that appears to only just be occuring to US top brass…

Clive Robinson • November 25, 2015 6:00 PM

@ Markus Ottela,

… there isn’t excess space for sophisticated malware.

It’s a very important security mitigation that few appear to pick up on, with all types of memory.

It was one of the primary security considerations behind the idea behind the design of the “cells” in my “prison” design.

A Point that is lost on most people is it actually takes more code “to emulate a process” than “the process it’s self does” (unless the process has been written very inefficiently). So the memory required by sophisticated “hiding” malware is not just the “payload” but the “process emulation” as well. Which adds significantly to the memory required. If you thus limit via hardware –MMU etc– to just that the process requires then the malware has nowhere to hide…

Clive Robinson • November 26, 2015 5:06 AM

@ Nick P, Markus Ottela,

For the latter [optical media], there’s even shredders designed with them in mind.

Optical media is quite problematic when it comes to “effective” destruction of data upon it. The problems arise from the “data density”, “data format and error correction” and the media “physical properties”.

When optical media first appeared it had some real issues, some was the assumption that the way magnetic media was used would apply. Thus the data storage format and error correction was not appropriate and caused high error rates. When combined with the fact that the original materials were to easily damaged extream care had to be practiced, similar to that of the early “disk pack” hard drives. Thus the drive and disk mechanics made the technology prohibitively expensive.

Thus a re-think was required, and it came not from the need to store data but audio recordings to replace “vinyl”. The physical data format was changed from tracks and sectors to a continuous spiral considerable dispersed data and error correction was added and a cheap to manufacture “pressed” format on a very physicaly tough material (polycarbonate) was used. Thus the audio CD became very reliable, so much so you could drill 2-4mm holes in many places and it would still play without audio fault.

The audio CD was ignored by the computer industry who were still fixated on replacing magnetic hard drives. Untill a small relatively unknown company with a need to send out immense static data bases realised the auduo CD potential. I worked for the company (SilverPlatter) and met and talked to some of those involved. They had a lot of trouble getting the computer industry interested, as SilverPlatter’s business was seen as very very niche. However to cut a long story short they pushed the technology and got it into many universities and scientific organisations, they also decided that it would be desirable that their drives were not just for data but could also do “audio” both in digital and analog form, and developed drivers for both. When they had it up and running Micro$haft realised it would solve their “bloatware distribution” issues and did their usuall “embrace, extend, capture, exclude” trick.

However MS hardly changed the work SilverPlatter had done, other than add a few things based on other work to make them more MS friendly in the areas of “booting” and “file system overlays”.

The result the CDROM is tough durable and the data is replicated and error corrected and thus smeared across large areas of the disk. All of which makes data destruction difficult at best.

Polycarbonate is quite durable and it’s surface difficult to destroy, thus even mechanically crunching the disk to rice grain particle size still leaves many grains with the same data in readable format for those with around 40,000USD of second hand lab equipment, or if a little inventive with about a quater of a man years work and about 1500USD of COTS equipment and hand made modifications.

Similar issues with “grinding” one or both surfaces, unless the actual impressed tracks are fully ground down far enough that pressing stresses are also removed data can be recovered, simply by carefull polishing and DSP tricks on the analog signals from a good quality drive.

Whilst you will by no means get all the data by either method, you will get a lot of fragments, which means that it is essential that all user data be encrypted as well as file system data, and importantly also at the lower level container level as well. It’s this last level which is not realy practical with COTS or professional recording equipment.

Floppy disks however can be easily destroyed, you can use a strong degausing magnet, but simpler is just to twist the plastic case, pull out the metal hub and thin magneticaly coated plastic disk, and melt/burn the plastic over a cigarette lighter, whilst avoiding breathing in the fumes.

Clive Robinson • November 26, 2015 9:54 AM

@ Nick P,

why not tape drives/cassettes for the increased capacity with same advantage that you can degauss and destroy the tape?

Whilst degaussing is nice, you have to also think about what happens when “they cut the power”?

Floppies unlike CDROMs and some tape drives still open at the push of a button due to their simple mechanical design. Which is a major advantage when it comes to quick destruction of the media.

Further destroying a floppy with a simple cigarette lighter is the work of thirty seconds at most. Even if you could get the tape out quickly it’s actually quite difficult to destroy. As you may have noticed they are much more robustly made often requiring you ti physically stamp/jump on them several times before you can get at the tape. Then your problems start, burning a tightly wound spool of tape is a bit like trying to burn a tightly rolled up newspaper, it’s dificult and takes time as there is to little air for the tape to actually burn, and not enough heat from a lighter to ensure it realy melts into an unrecoverable blob with some tapes.

The soulution is to keep the primary / master key on a piece of permanginate or nitrate washed “cigarette” paper and secondary keys on a floppy and chain encrypted data on tape or other storage. It the paper is potassium permanginate washed, you don’t even need a lighter, gust a quick wipe with glycerin and drop in the waste basket 😉