The problem is, that if the VM writer tries to take every possible method to make the execution time similar (e.g. make privileged instructions run as fast as non-privileged instructions), it has to slow the faster ones down. Suddenly, even your grandpa will notice something is wrong. The most insane method would be a VM based on a full-blown, cycle-accurate simulator, but that will be horribly slow.
Instead, what I think is it's not *impossible* to detect, but it's *difficult* to detect, because the VM detector is going to need a very very very long checklist to determine whether it is running on a VM or not. To be sure, it must check every possible privilegd instruction's timing, check the system memory's contents using various workarounds (such as DMA), and etc. etc.
Yes, "limited server side resources" is a problem. If you apply the FPS server model, you aren't dealing with 30-60 objects : you are dealing with 300-600 objects. A typical 1-to-1 starcraft game tends to go all the way up to 300+ objects, and if you are thinking of 3-on-3 or something close to that, you are screwed. Now, multiply the 300 objects by the number of battle.net players. Suddenly, Blizzard needs to assign one machine per game (that is, approximately six players) and still cannot be sure if they can manage to do it. Plus, the players won't even be possible to play without insane amount of bandwidth and network stability.
What's worse, is that the reason the FPS/MMORPG server model works for 30-60 players is that the server doesn't need to distribute all the players' movements to all players. Only a small number of players (around a dozen) will be sufficient. However, RTS games tend to have a large number of objects within vision. Thus, the client side needs to track the movement of about 200 objects even when the total objects on the game is around 300.
I suspect the method that Starcraft (and I believe, Warcraft 3, too) uses is by transmitting every player's mouse and keyboard behavior to each other. Each player runs its own world with the same input. As long as the game is completely deterministic (with all random number generators seeded with the same number), they have the same behavior. If there is any inconsistiency, it's either software bug (which sometimes did happen on corner cases), or some kind of hack, and the player gets disconnected. Since it's unlikely to have a player generate more than ten operations per second, the network overhead is low.
One possible anti-hack solution is to cross-check critical data of the game, for example, once every five seconds. These critical data can be anything - unit position, amount of resources, fog-of-war, etc. If any player disagree, the game is automatically canceled and the replay file (which consists of all players' key/mouse movement) is submitted to Blizzard for further analysis whether its a bug or a hack. Combined with other memory-checking or whatsoever blah blah solution, strict banning policies, and some FUD, I guess it will be enough to scare away most people.
Remember even the maphack people are playing for fun, not for profit. If playing for profit, you shouldn't be playing on the net anyway.
Here in Korea, the major ISP actually tried to implement some policy similar to this (although with clearer terms), but they somehow failed to do so. I guess the government didn't approve the new terms, due to the complaints from the media industry so that the ISPs can misuse the terms to lock out competitors and force them to use their own (inferior) service.
There simply are too many bandwidth-hogging services. My mom and dad watches IPTV (real-time h.264 full-HD, 3000kbps minimum, which translates to about 1.3GB for an hour of TV) from a competitor of our ISP, and I also use many video services, which are usually around 2000kbps/sec. Even if we don't abuse anything, I guess our monthly usage would be somewhat around 200GB.
Anyway, it simply looks stupid to specify the policy in terms of 'number of songs'. sad.
AFAIK, the problem of these kind of scheduling is that the priority cannot be determined as a single number.
For example, mp3 players require high priority only when the output buffer is low. When the buffer is nearly full, it doesn't need to have high priority, and can be kept idle. If the scheduler simply assumes high priority, then the mp3 player will be scheduled frequently, even when it only needs a small fraction of the quantum (around 10ms). This will increase the number of context switches, so you have an inefficient system which simply wastes so many cycles moving around threads. As many of you know, context switching is expensive, not because of the extra code that needs to be run, but because it unloads/loads cache, which adds more traffic on the memory side.
Of course, there can be other tricks to relieve this problem by adding neat tricks on the application (which certainly isn't a good idea, since it dependes on other threads and IO workload) or the scheduler (which means the scheduler needs is more information about the application). In the case of multimedia streaming, the information is 'the deadline it must meet' - for example, a pre-defined or measured amount of computation power per sec. Thus, something other than thread priority is needed to be fed to the scheduler. In the case of Vista, MMCSS is what handles it.
At least, according to the article, MMCSS *cripples* network performance for giving mulimedia applications a little more performance. The problem of the current MMCSS implementation is that it *cripples* network performance even when it isn't necessary, to a ridiculous extent.
Yes, the problem is with the newer OS - but the problem isn't a flawed concept. It's simply poor implementation.
Although I think some kind of multimedia-aware scheduler is absolutely necessary, it simply looks insane to hard code the network device to drop the rate to 10k packets/sec. This makes me sort of feel like they added MMCSS on the last moment of vista design, or they simply didn't think hard enough.
There can be two possible solutions 1) dynamically change the packet rate depending on the multimedia workload/processor/whatsoever 2) dropping network packets not on the NDIS, but on the TCP/IP driver, which starts dropping packets when multimedia processes runs out of processing power
Yes, I understand that your PC plays mp3s without any problem even on a GB ethernet connection. However, I don't think the MMCSS is a flawed concept/unneeded feature/whatsoever. The MMCSS is for improving multimedia performance on EXTREMELY heavily-loaded processors. I use XP, and my PC is occasionally heavily loaded with a dozen threads, and in those cases I occasionally experience glitches. Thus, I have to manually adjust thread priorities, but it's annoying anyway.
The problem is a flawed implementation - nothing else.
ps : I wish there was some method to reserve some processing power for launching taskmgr, so that I can kill problematic processes with ease. Any suggestions?
Not exactly. They said it is a microcode update, which updates the processor, not the operating system.
The reason it only affects Microsoft Windows may be because only Windows triggers that processor bug, and other operating systems doesn't. Of course, there can be software workarounds (rewrite Windows to bypass the bug), but why rewrite code when Intel already released their 'patch' for their microprocessor?
FYI, 'microcode' is simple code (not x86, an internal format) that runs inside the processor. Normally, simple instructions gets translated into a single 'internal operation', but if the instruction is complex, the internal 'microcode engine' kicks in, and runs the microcode that mimics the instruction's behavior. What's amazing, is that a microcode fix can eliminate so many kind of bugs (sometimes, by disabling hardwired logic and replacing it with microcode).
Actually, Both Intel and AMD releases microcode updates frequently, fixing thousands of bugs even after product release.
PS : I don't work for any of the companies mentioned above, so details may vary.
Sounds very expensive to me. Here in Korea the normal price for 1 sms is around 3 cents, but no operators offer flatrate, due to its abuse by spammers.
At first, in Korea, all major carriers had provided flat rate plans, but once cell phone spammers started to abuse them (custom SMS spamming software + PC + flat rate plan = unlimited spamming capability), they quickly removed it from their plan list. Seriously, I used to recieve 20+ SMS spams a day on the worst ages, but once carriers started to provide spam filtering (free of charge, can disable any time), the spam rate dropped to less than one per week.
However, there still are service plans that even provide 1000+ free messages per month, and it seems to be enough for most people.
Anyway, flatrate seems to be troublesome, and it seems to be (sort of) surprising that US carriers have those kind of rates.
How in world did Zonk 'imagine' that Samsung signed patent cross-licensing because of Linux? Remember that Samsung is a fairly large company, and I remember that most of the software code Samsung uses isn't based on open-source anyway.
Come on. They are simiply trying to minimize business risk, not trying to bash/crush/destroy Linux. Nothing else.
Yes, this is Slashdot, so anybody who does anything with Microsoft will be evil. Oh. I see. Never mind.
(btw, I believe that both parties would be interested in patents related to video/audio stuff, not something like operating systems.)
In an article that I can't remember where it's from, I saw that it's black simply for marketing 'coolness'. The actual products that will be delivered to customers will be white.
The Korean governmnent isn't telling the people to wait until Vista's security policy changes. It's the other way around - the Korean government is telling them to wait until third party vendors (including the Korean governmnent itself) adopt to the new Vista security model.
I actually hope Vista adopts a REALLY harsh security model, so that the ill-designed, ActiveX-cluttered website designers have no choice but to design their site so that they have no administrator privilege.
Though I read the article, I could no find any details on the patent. However, the tone of the article seems it is referring to 'automatic input mode switching' method.
Very often, people type Korean while the input mode is in English, and vice versa. What the software does, is that it detects the context of the typing sequence, and figure out whether it needs automatic mode change or not. For example, detecting invalid Korean is simple -- the software simply seeks for invalid typing sequences, since each Korean syllable contains of a sequence of consonant - vowel - consonant (while the last consonant is optional). Detecting invalid English sequences will be a little bit more difficult, but it seems to be possible using clever techniques combined with large dictionaries.
If the detection algorithm is crappy, typing becomes a nightmare, since it transforms correct English into incorrect Korean typing (or vice versa). However, in these cases, the 'auto-mode-change' can be turned off, or custom words can be added into the dictionary. The two most popular word processing softwares (MS Word and Hangul http://en.wikipedia.org/wiki/Hangul_(word_processo r)) both has a good implementation.
The wierd thing is that I could not find any source of this news on most major Korean sites.
I'm on my first year of Ph.D course, and there is another good reason other than ones mentioned here.
In Korea, everybody men must be on millitary service for a certain period (24 months for army, a little longer for navy or air force). However, with a MS degree on many engineering/science degrees, there are many alternative ways to do the millitary service.
Research positions on goverment agencies related to defence, corporates related to defence, research positions on universities, et cetra. Well paid, do interesting work, (at least more interesting than patrolling around the DMZ) developing more skills on your field, plus, you can continue your Ph.D course while on millitary service.
The only downside of this is that you can only stay in Korea, need a MS, and have a longer service period (3 years). Another downside is the risk - the number of seats are quite limited, and if you fail to get the position, or worse, fail to get your MS degree, you eventually get forced to join the army, no matter what kind of plans you had.
Yes, I'm doing my military service as a research staff in a university.
However, in the real world, what I suspect is that this kind of search will be done with some kind of warrant - I don't think FedEx will allow MPAA to search their stuff without any warrant.
Actually, this is exactly what happened in Korea. Proscecutors + BSA + warrants raiding offices (end users) searching for pirated software. The result? massive fines which made many small companies out of business.
ARM already dominates its own market - embedded processors.
For example, I already have an wireless AP, a cellphone, and an MP3 player that is known to use an ARM procesor. Adding up many more 'unknown' gadgets I have, I'd bet that I myself already have a dozen of these.
Go to ARM's homepage, and find a short list of gadgets that use ARM. http://www.arm.com/
Although it may look like it requires a lot of overhead to run a VM monitor as a rootkit, it doesn't look that complicated, since the difference between a rootkit VMM and a real product (such as VMWare) has these kind of differences
1) VMWare needs to emulate the hardware to provide a standard interface, while the rootkit simply may bypass most operations to the hardware, and process the ones that they need to 2) VMWare needs to share the hardware between various other VMs and host machines, while the rootkit doesn't need to (the host's behavior is completely predictable, and there is only one VM)
Since the malicious VMM doesn't need to do much work (other than protecting itself and do some malicious work), the binary doesn't need to be so large. Since it bypasses most operations, the overhead won't be so large either (I guess, less than 1%) Moreover, it doesn't need to support most hardware (because it will bypass all operations anyway).
Although TCPM may look like a solution, I think hardware-based virus scanners may be sufficient to kick off VMM-based rootkits. Make it act as a firewall between the disk and the mobo. Monitoring the disk traffic, plus the BIOS flash seems to be enough.
1. Capacity : There are many people who can live with 5GB rather than 500GB. 2. Lifespan : Think how many times you would erase a part of the flash in a year.
At least, I would buy one if it provides 20GB @ $300. Better performance, better reliability, lower power, all at the cost of an tenfold size decrease, but I don't need that must space on my laptop. (I would still need a high-capacity hard drive on my PC, though.)
First, I'm sad to say that ISO9660 is designed to be read-only, thus lacks features that are essential for rewriting stuff. Yes, there can be some ways to make it writable, but it would require the system to be extremely inefficient, and/or have many restrictions on possible operations. (for example, you may have a filesystem that won't return erased space to free space, unless you reformat the filesystem. Maybe possible for digital cameras, but I'd rather pay 25 cents.)
I think the best move that the 'potential' victims to do is to file a lawsuit against Microsoft for invalidating the patent. However, I'm not sure if there would be any company which would be happy to take the risk.
Not only subway lines, but in Korea, cell coverage is almost perfect.
What I heard from a cell carrier executive (on a seminar), is that they actually developed repeaters that costs about $3000, and installed 250K of them. yes, two hundred and fifty thousand of them. (It was last year when I heard this, so it should be even cheaper now) Multiply it by three for all three major telcos, and that's approximately 800k repeaters.
You find cell coverage is somewhat poor in your building, then call the carriers, and they will install devices for you. Yes, all for free. Even if there are only about 10 users on that area.
First, you should know that the other companies, such as Cowon (never heard of?) or Iriver (maybe?) doesn't buy something like 10k flashes, they buy MILLIONS of them. Although that's still a lot less than Apple's product, that's still an awful lot of chips. You should know that iPod's market share in Korea is completely disappointing, http://www.cdfreaks.com/news/12253 and companies like iRiver still sells millions of players every year.
The Korean manufacturers claim that Samsung never sell flash memory chips directly to them, no matter how many chips they buy. (unlike Apple, which gets flashes directly from Samsung) Adding an additional layer on sales means more cost. Additionally, they claim that Samsung never ever sells 2Gig chips to any Korean mp3 company, no matter how much they pay. Though I don't know much about Korean law, but they claim that it's against the Korean law (Looks like some antitrust/fair trade bill).
The second problem is that there is a serious flash shortage (probably due to Apple's nano), and many companies (especially smaller ones) claim that they are treated unfairly, compared to the larger corporates.
What the Korean mp3 companies, plus many other people claims is that Samsung's strategey to sell flash chips with near zero margin, or even with loss, is:
1) Kill hard disk companies, such as Toshiba or Hitachi, that sold disks to Apple. Note that Samsung also sells hard disks, and pressurizing competitors HARD will definitely be good for Samsung's hard disk business.
2) Suggest an appealing price to Apple, and lock them to flash-only players,
3) Kill the local mp3 business, and kill competition in Korea. That will help Samsung's mp3 business (plus, the cellphones with mp3 player capability). After all competition is gone, they can deal with Apple by stop providing flash with huge discounts.
Although none of these claims are proven to be true yet, it's clear that the Samsung-Apple deal pressurizes both hard disk manufacturers and mp3 player manufactures.
As the prothread homepage says, it's for extremely small embedded systems, where there are no operating systems, with tiny amount of memory (You can't use DRAMs on systems that cost something less than $1). Want to use threads on those kind of systems, you have no choice.
Another advantage is its portability. Small embedded systems, whether they have operating systems or not, usually can't support some fully-blown threading standard. Those operating systems seem to implement some kind of 'specially tuned' thread APIs.
Using these kind of threads on a full-blown PC (or servers) would have almost no benefit. However, in the embedded software engineer's perspective, it's great to see a ultra-lightweight thread library without any platform-dependent code.
Slashdot Mirror
The problem is, that if the VM writer tries to take every possible method to make the execution time similar (e.g. make privileged instructions run as fast as non-privileged instructions), it has to slow the faster ones down. Suddenly, even your grandpa will notice something is wrong. The most insane method would be a VM based on a full-blown, cycle-accurate simulator, but that will be horribly slow.
Instead, what I think is it's not *impossible* to detect, but it's *difficult* to detect, because the VM detector is going to need a very very very long checklist to determine whether it is running on a VM or not. To be sure, it must check every possible privilegd instruction's timing, check the system memory's contents using various workarounds (such as DMA), and etc. etc.
Yes, "limited server side resources" is a problem. If you apply the FPS server model, you aren't dealing with 30-60 objects : you are dealing with 300-600 objects. A typical 1-to-1 starcraft game tends to go all the way up to 300+ objects, and if you are thinking of 3-on-3 or something close to that, you are screwed. Now, multiply the 300 objects by the number of battle.net players. Suddenly, Blizzard needs to assign one machine per game (that is, approximately six players) and still cannot be sure if they can manage to do it. Plus, the players won't even be possible to play without insane amount of bandwidth and network stability.
What's worse, is that the reason the FPS/MMORPG server model works for 30-60 players is that the server doesn't need to distribute all the players' movements to all players. Only a small number of players (around a dozen) will be sufficient. However, RTS games tend to have a large number of objects within vision. Thus, the client side needs to track the movement of about 200 objects even when the total objects on the game is around 300.
I suspect the method that Starcraft (and I believe, Warcraft 3, too) uses is by transmitting every player's mouse and keyboard behavior to each other. Each player runs its own world with the same input. As long as the game is completely deterministic (with all random number generators seeded with the same number), they have the same behavior. If there is any inconsistiency, it's either software bug (which sometimes did happen on corner cases), or some kind of hack, and the player gets disconnected. Since it's unlikely to have a player generate more than ten operations per second, the network overhead is low.
One possible anti-hack solution is to cross-check critical data of the game, for example, once every five seconds. These critical data can be anything - unit position, amount of resources, fog-of-war, etc. If any player disagree, the game is automatically canceled and the replay file (which consists of all players' key/mouse movement) is submitted to Blizzard for further analysis whether its a bug or a hack. Combined with other memory-checking or whatsoever blah blah solution, strict banning policies, and some FUD, I guess it will be enough to scare away most people.
Remember even the maphack people are playing for fun, not for profit. If playing for profit, you shouldn't be playing on the net anyway.
Here in Korea, the major ISP actually tried to implement some policy similar to this (although with clearer terms), but they somehow failed to do so. I guess the government didn't approve the new terms, due to the complaints from the media industry so that the ISPs can misuse the terms to lock out competitors and force them to use their own (inferior) service.
There simply are too many bandwidth-hogging services. My mom and dad watches IPTV (real-time h.264 full-HD, 3000kbps minimum, which translates to about 1.3GB for an hour of TV) from a competitor of our ISP, and I also use many video services, which are usually around 2000kbps/sec. Even if we don't abuse anything, I guess our monthly usage would be somewhat around 200GB.
Anyway, it simply looks stupid to specify the policy in terms of 'number of songs'. sad.
AFAIK, the problem of these kind of scheduling is that the priority cannot be determined as a single number.
For example, mp3 players require high priority only when the output buffer is low. When the buffer is nearly full, it doesn't need to have high priority, and can be kept idle. If the scheduler simply assumes high priority, then the mp3 player will be scheduled frequently, even when it only needs a small fraction of the quantum (around 10ms). This will increase the number of context switches, so you have an inefficient system which simply wastes so many cycles moving around threads. As many of you know, context switching is expensive, not because of the extra code that needs to be run, but because it unloads/loads cache, which adds more traffic on the memory side.
Of course, there can be other tricks to relieve this problem by adding neat tricks on the application (which certainly isn't a good idea, since it dependes on other threads and IO workload) or the scheduler (which means the scheduler needs is more information about the application). In the case of multimedia streaming, the information is 'the deadline it must meet' - for example, a pre-defined or measured amount of computation power per sec. Thus, something other than thread priority is needed to be fed to the scheduler. In the case of Vista, MMCSS is what handles it.
At least, according to the article, MMCSS *cripples* network performance for giving mulimedia applications a little more performance. The problem of the current MMCSS implementation is that it *cripples* network performance even when it isn't necessary, to a ridiculous extent.
Yes, the problem is with the newer OS - but the problem isn't a flawed concept. It's simply poor implementation.
Although I think some kind of multimedia-aware scheduler is absolutely necessary, it simply looks insane to hard code the network device to drop the rate to 10k packets/sec. This makes me sort of feel like they added MMCSS on the last moment of vista design, or they simply didn't think hard enough.
There can be two possible solutions
1) dynamically change the packet rate depending on the multimedia workload/processor/whatsoever
2) dropping network packets not on the NDIS, but on the TCP/IP driver, which starts dropping packets when multimedia processes runs out of processing power
Yes, I understand that your PC plays mp3s without any problem even on a GB ethernet connection. However, I don't think the MMCSS is a flawed concept/unneeded feature/whatsoever. The MMCSS is for improving multimedia performance on EXTREMELY heavily-loaded processors. I use XP, and my PC is occasionally heavily loaded with a dozen threads, and in those cases I occasionally experience glitches. Thus, I have to manually adjust thread priorities, but it's annoying anyway.
The problem is a flawed implementation - nothing else.
ps : I wish there was some method to reserve some processing power for launching taskmgr, so that I can kill problematic processes with ease. Any suggestions?
Not exactly. They said it is a microcode update, which updates the processor, not the operating system.
The reason it only affects Microsoft Windows may be because only Windows triggers that processor bug, and other operating systems doesn't. Of course, there can be software workarounds (rewrite Windows to bypass the bug), but why rewrite code when Intel already released their 'patch' for their microprocessor?
FYI, 'microcode' is simple code (not x86, an internal format) that runs inside the processor. Normally, simple instructions gets translated into a single 'internal operation', but if the instruction is complex, the internal 'microcode engine' kicks in, and runs the microcode that mimics the instruction's behavior. What's amazing, is that a microcode fix can eliminate so many kind of bugs (sometimes, by disabling hardwired logic and replacing it with microcode).
Actually, Both Intel and AMD releases microcode updates frequently, fixing thousands of bugs even after product release.
PS : I don't work for any of the companies mentioned above, so details may vary.
Sounds very expensive to me. Here in Korea the normal price for 1 sms is around 3 cents, but no operators offer flatrate, due to its abuse by spammers.
At first, in Korea, all major carriers had provided flat rate plans, but once cell phone spammers started to abuse them (custom SMS spamming software + PC + flat rate plan = unlimited spamming capability), they quickly removed it from their plan list.
Seriously, I used to recieve 20+ SMS spams a day on the worst ages, but once carriers started to provide spam filtering (free of charge, can disable any time), the spam rate dropped to less than one per week.
However, there still are service plans that even provide 1000+ free messages per month, and it seems to be enough for most people.
Anyway, flatrate seems to be troublesome, and it seems to be (sort of) surprising that US carriers have those kind of rates.
How in world did Zonk 'imagine' that Samsung signed patent cross-licensing because of Linux? Remember that Samsung is a fairly large company, and I remember that most of the software code Samsung uses isn't based on open-source anyway.
Come on. They are simiply trying to minimize business risk, not trying to bash/crush/destroy Linux. Nothing else.
Yes, this is Slashdot, so anybody who does anything with Microsoft will be evil. Oh. I see. Never mind.
(btw, I believe that both parties would be interested in patents related to video/audio stuff, not something like operating systems.)
In an article that I can't remember where it's from, I saw that it's black simply for marketing 'coolness'.
The actual products that will be delivered to customers will be white.
Yup. There is a word processor called "Hangul".
t _info/office/hangul2007_info.php
http://www.haansoft.com/hnc5_0/haansoft_en/produc
The Korean governmnent isn't telling the people to wait until Vista's security policy changes.
It's the other way around - the Korean government is telling them to wait until third party vendors (including the Korean governmnent itself) adopt to the new Vista security model.
I actually hope Vista adopts a REALLY harsh security model, so that the ill-designed, ActiveX-cluttered website designers have no choice but to design their site so that they have no administrator privilege.
Though I read the article, I could no find any details on the patent. However, the tone of the article seems it is referring to 'automatic input mode switching' method.
o r)) both has a good implementation.
Very often, people type Korean while the input mode is in English, and vice versa. What the software does, is that it detects the context of the typing sequence, and figure out whether it needs automatic mode change or not. For example, detecting invalid Korean is simple -- the software simply seeks for invalid typing sequences, since each Korean syllable contains of a sequence of consonant - vowel - consonant (while the last consonant is optional).
Detecting invalid English sequences will be a little bit more difficult, but it seems to be possible using clever techniques combined with large dictionaries.
If the detection algorithm is crappy, typing becomes a nightmare, since it transforms correct English into incorrect Korean typing (or vice versa). However, in these cases, the 'auto-mode-change' can be turned off, or custom words can be added into the dictionary. The two most popular word processing softwares (MS Word and Hangul http://en.wikipedia.org/wiki/Hangul_(word_process
The wierd thing is that I could not find any source of this news on most major Korean sites.
You forgot to roast your coffee beans before grinding it.
I'm on my first year of Ph.D course, and there is another good reason other than ones mentioned here.
In Korea, everybody men must be on millitary service for a certain period (24 months for army, a little longer for navy or air force). However, with a MS degree on many engineering/science degrees, there are many alternative ways to do the millitary service.
Research positions on goverment agencies related to defence, corporates related to defence, research positions on universities, et cetra. Well paid, do interesting work, (at least more interesting than patrolling around the DMZ) developing more skills on your field, plus, you can continue your Ph.D course while on millitary service.
The only downside of this is that you can only stay in Korea, need a MS, and have a longer service period (3 years). Another downside is the risk - the number of seats are quite limited, and if you fail to get the position, or worse, fail to get your MS degree, you eventually get forced to join the army, no matter what kind of plans you had.
Yes, I'm doing my military service as a research staff in a university.
However, in the real world, what I suspect is that this kind of search will be done with some kind of warrant - I don't think FedEx will allow MPAA to search their stuff without any warrant.
Actually, this is exactly what happened in Korea. Proscecutors + BSA + warrants raiding offices (end users) searching for pirated software. The result? massive fines which made many small companies out of business.
ARM already dominates its own market - embedded processors.
For example, I already have an wireless AP, a cellphone, and an MP3 player that is known to use an ARM procesor. Adding up many more 'unknown' gadgets I have, I'd bet that I myself already have a dozen of these.
Go to ARM's homepage, and find a short list of gadgets that use ARM.
http://www.arm.com/
Although it may look like it requires a lot of overhead to run a VM monitor as a rootkit, it doesn't look that complicated, since the difference between a rootkit VMM and a real product (such as VMWare) has these kind of differences
1) VMWare needs to emulate the hardware to provide a standard interface, while the rootkit simply may bypass most operations to the hardware, and process the ones that they need to
2) VMWare needs to share the hardware between various other VMs and host machines, while the rootkit doesn't need to (the host's behavior is completely predictable, and there is only one VM)
Since the malicious VMM doesn't need to do much work (other than protecting itself and do some malicious work), the binary doesn't need to be so large. Since it bypasses most operations, the overhead won't be so large either (I guess, less than 1%) Moreover, it doesn't need to support most hardware (because it will bypass all operations anyway).
Although TCPM may look like a solution, I think hardware-based virus scanners may be sufficient to kick off VMM-based rootkits. Make it act as a firewall between the disk and the mobo. Monitoring the disk traffic, plus the BIOS flash seems to be enough.
1. Capacity : There are many people who can live with 5GB rather than 500GB.
2. Lifespan : Think how many times you would erase a part of the flash in a year.
At least, I would buy one if it provides 20GB @ $300. Better performance, better reliability, lower power, all at the cost of an tenfold size decrease, but I don't need that must space on my laptop.
(I would still need a high-capacity hard drive on my PC, though.)
I'd rather add that much RAM on my PC, since they are faster with a similar price.
You definitely should consider that RAM price is dropping as fast as flash memory price.
First, I'm sad to say that ISO9660 is designed to be read-only, thus lacks features that are essential for rewriting stuff. Yes, there can be some ways to make it writable, but it would require the system to be extremely inefficient, and/or have many restrictions on possible operations. (for example, you may have a filesystem that won't return erased space to free space, unless you reformat the filesystem. Maybe possible for digital cameras, but I'd rather pay 25 cents.)
I think the best move that the 'potential' victims to do is to file a lawsuit against Microsoft for invalidating the patent. However, I'm not sure if there would be any company which would be happy to take the risk.
Not only subway lines, but in Korea, cell coverage is almost perfect.
What I heard from a cell carrier executive (on a seminar), is that they actually developed repeaters that costs about $3000, and installed 250K of them. yes, two hundred and fifty thousand of them.
(It was last year when I heard this, so it should be even cheaper now)
Multiply it by three for all three major telcos, and that's approximately 800k repeaters.
You find cell coverage is somewhat poor in your building, then call the carriers, and they will install devices for you. Yes, all for free. Even if there are only about 10 users on that area.
First, you should know that the other companies, such as Cowon (never heard of?) or Iriver (maybe?) doesn't buy something like 10k flashes, they buy MILLIONS of them. Although that's still a lot less than Apple's product, that's still an awful lot of chips. You should know that iPod's market share in Korea is completely disappointing, http://www.cdfreaks.com/news/12253
and companies like iRiver still sells millions of players every year.
The Korean manufacturers claim that Samsung never sell flash memory chips directly to them, no matter how many chips they buy. (unlike Apple, which gets flashes directly from Samsung) Adding an additional layer on sales means more cost. Additionally, they claim that Samsung never ever sells 2Gig chips to any Korean mp3 company, no matter how much they pay. Though I don't know much about Korean law, but they claim that it's against the Korean law (Looks like some antitrust/fair trade bill).
The second problem is that there is a serious flash shortage (probably due to Apple's nano), and many companies (especially smaller ones) claim that they are treated unfairly, compared to the larger corporates.
What the Korean mp3 companies, plus many other people claims is that Samsung's strategey to sell flash chips with near zero margin, or even with loss, is:
1) Kill hard disk companies, such as Toshiba or Hitachi, that sold disks to Apple. Note that Samsung also sells hard disks, and pressurizing competitors HARD will definitely be good for Samsung's hard disk business.
2) Suggest an appealing price to Apple, and lock them to flash-only players,
3) Kill the local mp3 business, and kill competition in Korea. That will help Samsung's mp3 business (plus, the cellphones with mp3 player capability). After all competition is gone, they can deal with Apple by stop providing flash with huge discounts.
Although none of these claims are proven to be true yet, it's clear that the Samsung-Apple deal pressurizes both hard disk manufacturers and mp3 player manufactures.
As the prothread homepage says, it's for extremely small embedded systems, where there are no operating systems, with tiny amount of memory (You can't use DRAMs on systems that cost something less than $1). Want to use threads on those kind of systems, you have no choice.
Another advantage is its portability. Small embedded systems, whether they have operating systems or not, usually can't support some fully-blown threading standard. Those operating systems seem to implement some kind of 'specially tuned' thread APIs.
Using these kind of threads on a full-blown PC (or servers) would have almost no benefit. However, in the embedded software engineer's perspective, it's great to see a ultra-lightweight thread library without any platform-dependent code.
You forgot Old Koreans!