Naked Science Forum
Non Life Sciences => Technology => Topic started by: Titanscape on 04/04/2010 18:13:01
-
Hi again, I am well intentioned an interested in the US and Russian Federation computer systems, also it just comes to mind about the Indian and Pakistani equivalents. What happens if by the Internet, a hacker, such as a terrorist or wicked Chinese hacker triggers off a nuclear strike? The consequences would be terrible, right?
Would it be a good idea if the computers attached to the missiles, was not accessible from th Internet at all?
A computer system using instead something like a UNIX based OS which is secret and has no systems familiar to hackers current experience. And computers with next to no HDD, but only a solid, non re-writable ROM computer memory for running it's programs, and these would be written in a Lab on the secret OS also without access to the Internet, just to be sure. Then burned onto the ROM. The ROM being non CD or DVD but a unique government ROM, perhaps a LP size DVD.
Each time a program alteration is made another ROM is burned.
What do you think?
-
India and Pakistan and China have nuclear weapons, there might be a danger of a terrorist or other sourced hack into the missile computers and a strike from error, leading to a nuclear war, world scale. Is this true?
What need be done to avoid this?
I did a thread in technology similar to this.
-
I don't think the systems are based on Windows :-) The "Blue screen of death" may not be such an exaggerated phrase if they were. I don't know what OS the military use but it would not surprise me if it were UNIX based. I don't think it likely that these systems are easy to hack to actually fire missiles. I think where hacks that have been done, the people caught were given very severe punishments when they probably accessed top secret data like the next day's lunch menu. As far as I know, serious operations can only be started by multiple "manual" verification via hardware.
-
The control computers probably run a custom machine code, and are totally separate from the outside world.
There is no need for "remote control" by any other method than a fixed line connecting each and every centre together, with heavy encryption and a very limited set of codes it understands after decoding. It will not connect in any form with the switched network, probably having a dedicated microwave relay system and buried line backups to relay commands and test data. Military computers tend to be based on 1960's era minicomputers and mainframes, and probably are secure because almost nobody uses the programming languages anymore, as well as being written from the outset to be paranoid about security and failing safely in case of error. Backup would be a printer producing paper records that cannot be altered, and the programs would probably be entered via punched tape. In most cases input will be via key switches, and the operators generally will not be the same group who maintain the system, and will be physically unable to access anything of the internals.
Just think that in these arenas the biggest problem is going to be spare parts, as no manufacturer nowadays makes eg a 4004 RAM chip, or a 2708 EPROM, let alone core planes, and I have used equipment that used all of these, although the failure rate on those parts was so close to nil as to be undetectable. Compare to your modern throwaway phone that dies after 3 years forcing you to buy new.
The Russians have it worse, they copied US technology and now the factories that made parts are now literally in different countries, and many have closed down, or are now making different products. They also used a lot of thermionic valves, and spares are now poor quality as the batches are smaller and the artisans who developed the hidden knowledge on assembly are aging or dead.
-
Don't you mean the Intel 4004 microprocessor?
-
I think nuclear weapon systems are designed to require both software(which would always have potential at least to be "hacked") and hardware(un-hackable) in order to fire the weapon.
I think that in order to fire a nuclear ICBM, a group would have to both be able to get past some pretty intense encryption, and be able to physically take over a nuclear facility.
On the other hand, the more prevalent nuclear weapons and ICBM's become, the more danger there is that they will fall into the wrong hands.
-
Have a read of:
http://en.wikipedia.org/wiki/Two-man_rule#Nuclear_weapons (http://en.wikipedia.org/wiki/Two-man_rule#Nuclear_weapons)
which describes how a minimum of two people are required to activate their launch keys within a very short period of time to prevent the launch by a single person alone. I also believe, although it is not explicitly mentioned there, that the hardware backup systems referred to operate by ensuring the systems critical to the launch are isolated from all other systems until the authorisation and simultaneous switching of the launch keys have been correctly performed by people and that these systems cannot be switched automatically or remotely. Thus, it should not be possible to 'hack' into the necessary systems because there is no connection to those systems until the point immediately prior to a previously authorised launch.
Having said that though, note that the article mentions 'Single Survivor' procedures. Also have a read of:
http://en.wikipedia.org/wiki/Dead_Hand_(nuclear_war) (http://en.wikipedia.org/wiki/Dead_Hand_(nuclear_war))
which allows for automated launching. However, and as with the 'Single Survivor' procedures, I believe that to enable the Dead Hand facility the same degree of authorisation procedures as a 'manual' launch are required and must be performed, so once again it should not be possible to remotely connect to the relevant systems until they've already been made ready for an imminent launch.
[edited to remove redundant comment re the thread title]
-
Could one of the mods please move my response from the duplicate thread titled '?' in 'Just Chat' to this one - at least this one has a meaningful title.
In addition to what I've posted there I'll just add that the electronics systems used in the critical launch systems are unlikely to use a conventional Operating System (OS) at all, and certainly not a commercial one, and most certainly not Windows. Instead, the control software will run directly on the hardware, to both reduce the complexity and to increase the reliability (as no OS means you don't have to worry about bugs in that underlying OS and only have to concern yourself with the functional logic).
-
I have merged the similar topic from Chat into this thread.
-
Geezer, no, I mean a 4004 RAM chip. So obsolete it only appeared in pre 1978 data books and was listed there as " Obsolete, no equivalent available" in all cases.
Not to mention the little collection of MPQ 6502 transistor arrays I have. They were used as core plane drivers, along with a few resistors, diodes and capacitors.
Also MC44 diodes, we used 1N400x diodes as replacements. Some new boards came in (!!pricey!!) and they used the same approach. The joys of obsolete but darn near indestructibile equipment, though I took 2 weeks to change a PROM, blew all the fuses using a breadboard ( OK 16 4 bit words) and spent the rest of the time resoldering dry joints. Filament displays anybody?
-
SeanB - Thanks! I must have used them too, but I've obviously forgotten the number!
I used filament displays too. Seem to remember "upgrading" to a new fangled chip from GI that could drive them directly. Woopdedo!!
-
Look however at the old danger scenario, a comet hits Pakistan, however more likely, a hacker, maybe a terrorist hacks into Pakistani computers for his religion and sends off a missile to India. What about the Pakis computers and systems?
-
I have merged the similar topic from Chat into this thread.
Thanks Geezer.
-
I have consulted my 1977 Texas instruments data book and it makes no mention of a 4004 semi conductor memory chip the smallest are 1024 bit chips although I remember changing 4004 chips in the olde worlde Siemens computers.
-
I found this
http://www.multigame.com/DRAMs_low.html
but I doubt that it's an exhaustive list.
-
I believe that the 4004 was in fact a static as opposed to dynamic RAM
-
No, it was a serial ram chip................ Think of a ccd delay line, but digital. AFAIK remember it was around 400 12 bit words long. Nice little TO 100 can.
-
This is an old thread!! But since someone has revived it...
A serial RAM is usually called a shift register. There were plenty around in the 70's. On the other hand CCDs are essentially analog delay lines in the sense that they passed on (at each clock) an analog level in the form of a quantity of charge. They have a physically smaller cell than a static RAM though and a very simple construction allows a high packing density so they could be a practical way of storing data economically but you would have to keep them recirculating data (refreshed to a correct logic level) so that they went completely around before the data leaked away (typically 2ms over temperature).
It was in the early 70's that Intel brought out the first dynamic memories (based on a 3 transistor cell). These got to 1k bits with the 1103 but then were superceded by more compact 1 transistor cell based devices (4k bits at first) and competition from Mostek (4027) and others. Actually the cell was 1 transistor and a capacitor. There was a huge expansion at this time and this qickly changed to 16k (4116) and 64k which was just about state of the art production by 1980. Static memories are always around a factor of 4 behind in density because of the cell size (6 transistor cell) but did not need refreshing. I used to design these things!!
Like Geezer the only 4004 I remember is the Intel 4bit microprocessor and the only 6502 I remember was also a processor (8 bit). But there were lots of naming overlaps especially when you don't remember the preceding letters.
-
We bought a 1k x 16 bit "flip-chip" memory card from DEC (RIP). If I remember, it used 1103s.
What a piece of junk! It corrupted all the time. As it turned out, it wasn't only the Bee Gees that were going home to Massachusetts.
-
The theme from the movie "War Games" where a kid hacks into the Pentagon, and almost sets off a thermonuclear war.
I believe some of our defence systems are connected to a private intranet, rather than the internet, although undoubtedly there is some crossover, for example VPNs are often carried over standard phone lines.
One of the problems is that any system has to have a "trigger mechanism". So, hack the trigger mechanism, and the system fails. So, if it requires a phone call from the president with certain activation codes, then if the activation codes become known, the rest can be faked. Or, if it uses some kind of public/private key, then hack the encryption key, and the codes can be generated.
And, to be effective, any system has to both be accessible, and have fail-safes. So, if something happens to the President, then the Vice President has to be able to respond to a pending nuclear attack within a minute or so, wherever he is at, thus more entry points.
-
I wish to dispute the fact that Soviet era valves (Tubes) Were substandard, I was delighted to find they had gold plated pins which greatly reduced the need to bang on the top of the TV to make it work.
-
I have a hard time believing that a fire control system for a nuclear missile would have an internet connection. I once worked supporting some nuclear missile systems, and we had a TEMPEST lab (http://en.wikipedia.org/wiki/TEMPEST) just to crunch numbers.
-
It is my supposition that in any defense system, the weakest link is the personnel department.