Secure hardware

Does the choice in processor, or harddisk, or board make some difference in the security of a computer-system? It doesn't have to be much difference but would e.g. an Intel-system be slightly more secure than an AMD-system? Considering that modern hardware tends to come with some possible bugs and/or possible exploits, I could imagine that some hardware components would be less vulnerable than other, similar hardware components. So I wonder, which hardware brands are best-known for the least possible vulnerabilities and bugs? (And not just processors, btw!)

Well.. that depends on your definition of "secure". If your talking about hacking/exploiting a processor, mobo, video card, ram, PSU, etc... lol, then no it doesn't matter. If your talking about heat, that's different. What processor runs cooler, etc... But there has already been tons of discussions on that topic...

Back in the day BIOS virii were pretty common, but nowadays you got BIOS write protection, which means nothing can write to the BIOS unless your turn it off...

Not too sure about right now. You might want to do some googling (and a wiki :)) for Microsoft Palladium which was the first name of MS proposed secured system for both hardware and software.

http://www.epic.org/privacy/consumer...palladium.html
http://www.theregister.co.uk/2002/06...ts_it_vendors/

The info in the links are out of date butthey still give some indication about the way MS and partners were thinking. MS has backtracked on a lot since then.

I think it's been renamed as 'Trusted Computing' which the public spin being that only authorised code will run on the machine and spyware/viruses ect will be blocked within the hardware. Those with a more paranoid nature belive the systemn will be used to block the machine to only play DRM software/ hardware and media files.

Some of these ideas are becoming apparent in MS Vista
--or not. You've got me looking now.

http://www.theregister.co.uk/2005/09...g_controversy/

https://www.trustedcomputinggroup.org/home

There are some things you might wish to consider based on physical security.

A CD ROM/Floppy/USB port offers a route to get malware onto the system.

A case that can be opened can have other hardware added (eg ADSL modem to allow ecternal access) or the hard drive to be removed and cloned, allowing the system to analysed/craked at leisure

A computer that can be physically carried out of the room allows the system to be opened and hence the above applies

Etc, etc

Hi

Let me look at it from a slightly different angle. The question I ask
myself is, how do you evaluate the security a "system" provides?
"For this", criterions are specificied (ISO 15408). The EAL's are
described in an overview at IACS[1].

At a certain (high) level of assurance, see DoD 5200.28-STD[2] it holds that
"no computer system can be considered truly secure if the basic hardware
and software mechanisms that enforce the security policy are themselves
subject to unauthorized modification or subversion."

Unfortunately, I would like to know much more about the certification processes than I
do, hence I am careful with what I am saying. Anyway, the highest (I think)
assurance level certified operating system, STOP[3a,3b], only has achieved
this certification by the use of particular hardware!

In this context, hardware is very relevant to achieve high-assurance level
certification. I would be happy to listen to more qualified voices.


Cheers.


[1] http://www.cesg.gov.uk/site/iacs/ind...displayPage=13
[2] http://www.fas.org/irp/nsa/rainbow/std001.htm
[3a] http://niap.nist.gov/cc-scheme/st/ST_VID3012a.html
[3b] http://www.cygnacom.com/labs/pfSEL0181T3b.htm

I believe this would fall under assurance and good architecture, etc .......... how can you design good code following the criteria and then piss away all your potential assurance by not holding your hardware to the yardsticks of the industry?



(Orange book)
3.3.3 Assurance
3.3.3.1 Operational Assurance
3.3.3.1.1 System Architecture
3.3.3.1.2 System Integrity

Quote:

---

The TCB shall maintain a domain for its own execution
that protects it from external interference or tampering
(e.g., by modification of its code or data structures).
The TCB shall maintain process isolation through the
provision of distinct address spaces under its control.
The TCB shall be internally structured into well-defined
largely independent modules. It shall make effective use
of available hardware to separate those elements that are
protection-critical from those that are not. The TCB
modules shall be designed such that the principle of
least privilege is enforced. Features in hardware, such
as segmentation, shall be used to support logically
distinct storage objects with separate attributes
(namely: readable, writeable). The user interface to the
TCB shall be completely defined and all elements of the
TCB identified. The TCB shall be designed and structured
to use a complete, conceptually simple protection
mechanism with precisely defined semantics. This
mechanism shall play a central role in enforcing the
internal structuring of the TCB and the system. The
TCB shall incorporate significant use of layering,
abstraction and data hiding. Significant system
engineering shall be directed toward minimizing the
complexity of the TCB and excluding from
the TCB modules that are not protection-critical.
3.3.3.1.2 System Integrity
Hardware and/or software features shall be provided that
can be used to periodically validate the correct
operation of the on-site hardware and firmware elements
of the TCB

---

NCSC-TG-021(Purple Book)

Quote:

---

entails a global perspective of the entire system

---

TC-2 REFERENCE MONITOR PERSPECTIVE


Attached below is a simple yet comprehensive pdf

Once again, I have forgotten my glasses at school lol... but what I made out it's sounding like, I'm wrong and you guys are saying that there is a difference security wise between AMD and Intel for example... How is this?????

More features = More problems

Recently we had the fiasco over Intel's Hyperthreading and cache timings. This feature had the side effect of allowing you to explore what is in the cache as it is being used. That could make it somewhat less secure. But it isn't a major issue.


Another thing you may consider secure/insecure is the ability to perform un-intended hardware modifications to your processors.

Such as overclocking.

Intel tries their best to prevent overclocking, and recently one of their chipsets and processor combinations somehow enforced a tight FSB speed. But it was fairly quickly reverse-engineered by other chipset makers with their own compatible chipsets and such security measures were no longer possible. There are also some software modifications to many of their smaller processors that allow control over their CPU speeds.

AMD generally allowed many modifications but are now trying to prevent them...it used to be possible to cut your own speed modifications into the PCB, but now that is pretty difficult... There are some software modifications that allow you to change the speed the hardware operates at.

There are other processors that you can perform modifications to to overclock...including that in a TI-83 calculator, a Palm Pilot / PDA, etc. You can even overclock a clock, but that's not useful since it wouldn't tell the time anymore... You could also make a 12 hour clock with hour and minute hands out of a 5400RPM hard drive, but that's another story...


You pretty much can reprogram any hardware. Sometimes it isn't financially possible because you may need a fab or electron microscope to reverse engineer it, but it is theoretically do-able.

It probably doesn't help show how "insecure" it is as much as the more common things people here have mentioned... In some systems there are multiple redundant processors that check for errors on critical stuff, but that's very specialized and exorbantly expensive stuff...something you wouldn't see in your standard x86 processor workstation.

Quote:

---

Those with a more paranoid nature belive the systemn will be used to block the machine to only play DRM software/ hardware and media files.

---

And I am one of them. Today, hardware is irrelevant to computer security.
A processor will execute the instructions fed to it. A hard drive will save the file
you send to it. The monitor will display what it's told to display.

Some evil wannabe fascist monster<rcgreen starts foaming at mouth> idiots
in the business want to insert "security" into the firmware on devices like hard drives
and vid cards<rcgreen puts on tinfoil hat>, and it would be a baaaaaaaaaaad thing.

But the answer is no, hardware has(today) no effect on security AFAIK.
:cool:

Hardware has no effect on security? Okay, then how about network printers? And there's a lot more hardware that are almost computers all by themselves. Phones these days are also not that safe anymore as they used to be because some wireless virus that seems to be running around wildly.
But my main concern is about internal hardware though. One of the first Intel processors had a floating-point bug in them that would mess up some calculations. I can imagine that modern processors might have even more bugs. And maybe even bugs that would allow someone to break into the system or just break the system. And of course, not only the processor could be abused for this but perhaps also network cards, IDE cards, graphic-cards, etc.
It doesn't have to be about hackers getting access. It could also just be some nasty instruction to some piece of hardware that just kills the system until a reset will reincarnate it again.

I've read a book about computer-virusses that was printed in 1989. ("Computervirusses" By Data-Becker) It mentioned a virus that could actually damage a diskdrive by moving the drive head to a non-existing track. The hardware didn't check for illegal values thus the head started to bump against the case and thus could get damaged. These are the vulnerabilities that I am thinking about.

You going for the CIA description of information security:

Confidentiality, Integrity and Availability.

When you look at hardware security in terms of the CIA then reliablility becomes important in terms of integrity and availability and are easily measurable for hardware. Computer broke = no availability = security failure.

Hard drive corrupts = loss of data integrity = security failure
Network card losing packets = loss of integrity and availability.
Poorly designed CPU allowing overflows could = failure in confidentiality if the intruder gets root.


Think of an issue and think about how it affects CIA of a system.

Quote:

---

Hardware has no effect on security? Okay, then how about network printers? And there's a lot more hardware that are almost computers all by themselves. Phones these days are also not that safe anymore as they used to be because some wireless virus that seems to be running around wildly.

But my main concern is about internal hardware though

---

Exactly, and that's what he was talking about to, so why are you babbling on about stuff that is not part of the conversation??

I agree with rcgreen, I don't see how hardware really affects security these days... What's the point of making a hard drive spin so fast that it breaks?? Data loss?? Big deal, I got backups, you should to... Back in the day people used to do that for the hell of it, but now it's all about information. My last comment, that's what I was talking about, actually hacking into a computer and stealing information by exploiting internal hardware does not seem to be practical these days. The only useful thing you might be able to do is gain root by exploiting the CPU like aspman said... But even then, there are easier ways of gaining root....

Quote:

---

Originally posted here by The Duck
Exactly, and that's what he was talking about to, so why are you babbling on about stuff that is not part of the conversation??

---

Ehm... I did start this conversation... ;)

And of course the security of a network printer does matter. What if some trojan virus manages to infect not only computers but also printers? In the past this wasn't very likely but these days many hardware components are almost small computers all by themselves. A printer virus would of course be removed when the printer is turned off but many network printers are kept running for weeks, and maybe even longer.

The same for a graphical card. What if a worm can install itself in the memory of the graphics card? Again, it would be removed during a reset but of course you can always get infected again. Modern cards have at least 64 MB of memory for their own processing so there's enough room for a small worm to hang around.

And well, you might not be that worried about harddisks that break because some worm told them to because you have back-ups, but it is a costly operation to have them replaced.

Sure, having secure software is more important but that's just it. Software is getting so secure these days that hackers sooner or later will try different approaches to get access to sensitive information. A hardware attack might have less risks of detection. Some worm that installs itself in the video-RAM on start-up might not even be noticed. And it could collect some kinds of information until another application retrieves the collected information and sends it to the hacker. It would make some interesting spyware.

This question is a mixed bag...

At the standard consumer level, hardware is about performance and not security. This doesn't mean that hardware security controls are not available to consumers... just that we don't tend to use them. Software tends to have the requirement of universal functionality, so clearly vendor specific enhancements are largely ignored.

CPU security rings vary from vendor to vendor, unfortunately these don't tend to be effectively utilized by consumer operating systems. Some main boards offer memory and bus segregation, again these have little use at the consumer level.

Systems lock LOCK even go so far as to use a second processing system (System Independent Domain Enforcing Assured Reference Monitor | SIDEARM) to handle security related processing. If you think that is out there... don't forget about EM (TEMPEST) shielding and LED high/low water mark systems (to prevent the useful observation of LED emissions on networking equipment).

Hardware you will find at the consumer level includes NICs and hard drives with their own data encryption processors, but this doesn't seem to be quite what you are asking for. Nor is hardware with startup integrity checks like those found in high assurance systems and the new TCPA focused systems.

All insecurity issues stem from incomplete/ineffective security controls or a leverage failure. In this case the latter is important. To that end I would focus on hardware designed (and prolly marketed) with stability as the focus and not performance.

cheers,

catch

PS. sec_ware and !mitationRust are spot on, as is Aspman's second post the people here who think hardware has no impact on security clearly have had no exposure the higher assurance systems and fail to understand vulnerabilities work (see attached: section 4.2 a hardware fault can introduce a high degree of unpredictability to a process, even more than a software fault can... which in turn results in the greater likelihood of a vulnerability becoming exploitable)

http://niap.nist.gov/cc-scheme/index.html
http://www.radium.ncsc.mil/tpep/library/rainbow/
http://www.cesg.gov.uk/indexNS.cfm
http://www.cesg.gov.uk/site/iacs/ind...displayPage=13
http://csrc.nist.gov/
http://www.niscc.gov.uk/niscc/index-en.html
http://www.cert.org/
http://www.itsecurity.com/

Quote:

---

And of course the security of a network printer does matter. What if some trojan virus manages to infect not only computers but also printers? In the past this wasn't very likely but these days many hardware components are almost small computers all by themselves. A printer virus would of course be removed when the printer is turned off but many network printers are kept running for weeks, and maybe even longer.

---

Honestly, you insult my intelligence with this comment... Once again, you failed to realize what I was talking about. I was talking about hardware inside the PC tower... NOT nodes connected to a network...

Quote:

---

The same for a graphical card. What if a worm can install itself in the memory of the graphics card? Again, it would be removed during a reset but of course you can always get infected again. Modern cards have at least 64 MB of memory for their own processing so there's enough room for a small worm to hang around.

---

Since this was brought up, I'm curious, if this is possible, does anyone know what can be accomplished from this?? It is kind of interesting :)...



Now, correct me if I'm wrong, but software run from an OS like windows XP these days do not have actual access to hardware to prevent freezing and system instability... So wouldn't this mean that it would be extremely hard to accomplish exploiting a computer hardware wise?? Unless you were exploiting a windows 95/98 computer or had physical access to the computer... Am I right? Because honestly this is just something that popped in my head 2 seconds ago lol...

Quote:

---

Since this was brought up, I'm curious, if this is possible, does anyone know what can be accomplished from this?? It is kind of interesting

---

A better and more practical example would be NIC cache rather than Video RAM.

Quote:

---

Now, correct me if I'm wrong, but software run from an OS like windows XP these days do not have actual access to hardware to prevent freezing and system instability.. So wouldn't this mean that it would be extremely hard to accomplish exploiting a computer hardware wise??

---

Depends on the level that the software is run at... user space applications should not have this level of access... that doesn't mean they can't take advantage of hardware flaws. Duel processor cores might give rise to instances of synchronization/serialization errors, but let us not forget very simple flaws like filling up the audit partition to supress the the audit trail.

Hardware security is an important field, but for most of us it should amount to little more than selecting hardware that is unlikely to reach and unforseen state (non-confined failure) and to this end the hardware must be robust and stable above all else.

cheers,

catch

Hardware could be used to make a device secure, but there is a
trade off in functionality. Example, the write-protect on a floppy disk.
If the drive is designed so that the write-protect tab physically disables
the drive from writing, then your software can't override that,
any more than a remote "hacker" could override the fact that you
pulled the network cable out of your box to protect it from him.

There is a hierarchy of authority. If a hard drive had firmware onboard
in ROM, that filtered the type of data that could be written, no OS or other
software could defeat it. But if devices were designed with such limitations,
the unintended consequences would mean a lot of crappy hardware with
inane quirks that made them mostly unuseable.

General purpose hardware needs to be programmable, and adaptable
if only to reach a wide enough market to make money.

So, don't expect to see network cards on the market that are hard wired to
stop viruses, worms, or porn from reaching your computer.
Now if the DRM guys had their way...
:cool:

Well I couldn't stop researching this topic....... I ran into this road of different models and projects, some shady and I don't know how they measure up to the yardsticks of the industry or if they're even complete. XOM — a processor architecture, Cerium architecture, AEGIS (I know I thought of the radar too)

Quote:

---

the secure CPU. XOM and AEGIS

---

Quote:

---

physically tamperresistant processors to support tamper and copy evident computing. XOM and AEGIS do not to trust the operating.

---

Say what?

Implementing an Untrusted Operating System on Trusted Hardware
http://www-vlsi.stanford.edu/~lie/Pa...sp2003-sub.pdf

Quote:

---

Currently, there exist various initiatives that place the trust in
modern computing systems in a hardware component rather
in software only. In these systems, the applications don’t
trust the operating system to protect their data, and the operating
system does not trust the application to properly use
its resources. The result is that the interface that the operating
system exports to each application must change to
support the hardware security features, and some of the protection
aspects of the operating system must be moved into
the hardware.

---

Certifying Program Execution with Secure Processors
http://pdos.csail.mit.edu/~rtm/paper...um:hotos03.pdf

Quote:

---

Conclusion
This paper describes Cerium, a trusted computing architecture
that provides tamper-evident program execution.
Cerium uses a physically tamper-resistant CPU and a
-kernel to protect programs from each other and from
hardware attacks. Cerium reports what program is running
and what hardware and software environment surrounds
the program, so the a user can decide whether to
trust a program’s output.

---

When We talk about security what sort of security are we talking about:
Software, hardware, network.. all of the above?

with a PC, what are we after and where from are we protecting

1/ The complete denial of access to the PC's data, and its operational processes from:
a/ out side the room,
i/ outside the Systems Network
ii/ with network access
b/ inside the room, outside the systems Network, not in physical contact with the machine
c/ inside the room, on the systems Network, not in physical contact with the machine
i/ Guest/user access to network
ii/ Admin access to the network
d/ in physical contact with the machine, and with:
i/ only guest access to the system
ii/ with Admin access
e/ internal access to machine.. boot access

What level do we protect to..

just the data
Usage information.... user profile
the integrity of the software ie prevent user reverse engineering or cracking a program
Software profileing
prevention of running non certified or pirated software
Product branding (this is a MS computer piss off *nix ..etc)
Hardware locking.. modding/cracking

As already mentioned there are hardware criteria for security, as well as the various software many of us are familure.

Network security is another level

no matter how well secure your home, it isnt if you leave the key under the mat