The backdoor may be in the hardware
on
No Backdoor in Vista
·
· Score: 4, Insightful
Intel, HP, Dell, and Toshiba are including the Intelligent Platform Management Interface (IPMI) in many of their machines. IPMI is a "remote administration" tool embedded in the LAN hardware. It looks at UDP packets (on ports 663 and 664) and performs various commands on the target machine, completely independently of the operating system.
Here's the IPMI 2.0 rev 1 specification, a rather long PDF.
IPMI is very powerful. An IPMI session starts with a
Presence ping Any machine with IPMI hardware should answer a "presence ping" on UDP port 663. This identifies an IPMI-capable machine, and returns some vendor info. Anyone can send this.
This should work even if the machine is "turned off", as long as
it has standby power and is on a LAN.
Then, there's a challenge-response authentication sequence. More on this later.
Once you're in, here are some of the things you can do:
Power up the system. Power it down. Force a hard reset. Force a power cycle. Force a phony overtemperature condition (in hopes of getting a clean OS shutdown.).
Disable front panel controls (power off, reset, and standby buttons.)
Yes, that's really in the protocol. See section 28.6 of the specification. Remote control can also lock out the keyboard and blank the screen.
Set system boot options Or, what OS do we want to run today?
These include useful tools like "bypass user password".
There's more. Much more. Basically, you can remotely take over the machine, turn it on, inventory the hardware, load an operating system, boot it up, and talk to it.
IPMI's back channel can do more than this. With some help from the operating system (and yes, it's supported in Windows) you can do more remote administration functions.
This is great for administering your data center remotely. But it has darker implications.
Supposedly, most machines are shipped with IPMI mostly turned off, unavailable until a program is run on the machine to load in the keys that enable it. Supposedly.
Thus, all it takes for IPMI to be a "backdoor" is for a set of secret challenge/response keys to be preloaded into the IPMI chip. There's no way to read those keys. Short of taking the chip apart, gate by gate, there's no way to tell if there's a backdoor in there.
Or a set of keys might be loaded by the system integrator before
shipping the system. You can't tell. So that's where to put a backdoor, where no one can find it.
There's an open source, OpenIPMI, for sending IPMI commands on Sourceforge. Send "Presence pings" to the machines you have and see if they answer.
Marc Raibert did some great work on legged running machines back in the 1980s, at CMU and later at MIT. Then he left MIT to form a startup, Boston Dynamics. But Boston Dynamics didn't do much dynamics; they ended up doing kinematic animations for the military, like most games have now.
Work. Yes, a 35 hour week and 5 week vacations. An hour of work in France buys more stuff than anywhere else.
Energy. 80% of electricity in France comes from nuclear plants. Most of the rest is from hydroelectric plants. Cheapest electricity in Europe.
France exports electricity. Now that's energy independence.
Back in 1973, at the first "oil shock" of the Arab oil embargo, there was debate in France over what to do. The decision was made to go for energy independence. Unlike in the US, that decision was carried out. And now France is reaping the rewards. They don't have to fight wars for oil.
I went to a talk at Stanford where the lead architect of the Cell processor described that demo. That's running on a Cell processor which is actually doing the rendering. Sony's original plan for the PS3 worked that way, but they eventually put in a conventional NVidia chip.
So in the PS3, the Cell processors aren't doing the rendering.
The Cell should render about as well as everything else with a current NVidia part.
Flyovers are easy if you have enough RAM and a GPU. How much RAM did the demo rig have?
There are some advantages to operating on 48VDC, but unless you have the big battery room of a telephone central office, they're not that big.
What Rackable is really pushing is a system where AC to 48VDC conversion takes place in a unit at the top of the rack, and 48VDC is local to the rack. That, at least, simplifies the cable management.
One big advantage of 120/240VAC power distribution using US standards is that the connectors are standardized and reasonably idiot-proof. That is, if you can plug it in, you won't overload the power cord or the connector, and if you overload the branch circuit, a breaker will trip.
Outlet strips have circuit breakers, so you can't overload the cord to the outlet strip without a breaker trip.
There are NEMA standard power plugs for 15A, 20A, and 30A circuits, 120/240VAC, and single and three phase configurations. All this is standardized nationally and enforced by the National Electrical Code.
In contrast,
there are no simple standards for 48VDC. Most 48VDC gear has big screw terminals. There are no standard plugs and sockets. Somebody, preferably a licensed electrician, has to check all the data plates, add up the current loads, calculate voltage drops, size the wire and breakers, and torque the big screw terminals to the correct torque, using the correct lockwashers. Every time you add or change a load, somebody has to recheck the math. Errors can cause a fire. None of this is all that hard if you have basic power technician skills, but you can't just go casually plugging stuff in.
Although, since the development of the low-cost clamp-around DC ammeter, things have become easier in the DC world.
Sounds like a cross between a Nintendo DS and a Blackberry.
There are lots of attempts to come up with a new form factor and a market niche to go with it. It's getting wierd. Apple's new iPod speakers have the form factor of a ghetto blaster. (There's a strong resemblance between the Apple Hi-Fi and the 1984 Radio Shack boombox with a dock for a Sony Walkman.)
Tribe just did a site makeover. The main change is that you have to view more pages to find anything, thus resulting in more ad displays. Users hate the new format.
Myspace will probably get worse to become "competitive".
Social networking should be peer to peer. There's no piracy issue; it's all about ME. No problems with the RIAA or MPAA. There's an open source project for somebody.
Do we have a higher crime rate than 20 years ago? Yes.
Actually, we in the US don't. Crime is down. The aging population is the biggest driver of this.
Residential burglaries are way down because there's no market for used home electronics. Auto theft is down because cars are harder to steal.
And some of the big-city police departments, starting with NYC, finally figured out how to manage their cops.
2km of heavy metals. That would be at least as bad as the KT impact that wiped out the dinosaurs.
The discouraging thing is that we probably could, today, build automated spacecraft that could reach the asteroid and set off nuclear bombs to change the orbit. It would be profitable to nudge the thing into earth orbit. And if somebody screws up, we lose the planet.
It's Roland the Plogger again, this time hyping a NASA technology.
It's not much of a technology. That's called "structured light", and it's been used for years in industrial computer vision systems. It's one of the simplest ways to measure depth in an image.
It's not even new to law enforcement. Here's a PowerPoint presentation on using it to look at stamped logos in pills. This is from a 2004 conference in Dallas.
In the article, the author wrote that he was hired to write copy to plug "colloidial silver". Google's first hit for "colloidial silver" is currently Quackwatch.
"Colloidal Silver: Risk Without Benefit". "Long-term use of silver preparations can lead to argyria, a condition in which silver salts deposit in the skin, eyes, and internal organs, and the skin turns ashen-gray." They have pictures.
Remmber Bonded Spammer?, Ironport's attempt to do this? They sold that business off to ReturnPath, and it's still around, but not used much.
I divert Bonded Spammer mail to a separate folder. Let's see what's in there:
Spam from "alt.com", part of the Friendfinder spam network. Friendfinder had for a while the distinction of having the same IP address in both the SpamCop block list and the Bonded Sender allow list, both run by Ironport.
Spam from Movies Unlimited Video, but only one, and I did buy something from them once, a long time ago.
Spam from Bebo. Lots of it, all because Bebo doesn't use double opt in to verify e-mail addresses.
An "invitation" from LinkedIn.
A mail bounce from Google. (Google fixed that; for a while, they were signing their mail bounces, which made them a tool for denial of service attacks.)
That's Bonded Spammer's track record. Let's see how AOL does.
Which MMORPG has the biggest world population?
on
World of Queuecraft
·
· Score: 1
Which MMORPG has the biggest world population? That is, which seamless online 3D world has the most people, without "shards" or "instances" to divide up the world?
On heap management, is this just the low-fragmentation heap that Microsoft introduced in Windows Server 2003? That's just a homogeneous-block storage allocator, which isn't new. Hans Boehm's garbage collector uses one of those. If you round up allocations to the next standard size, you waste some RAM but reduce fragmentation. It's usually a win, but not a big one; it speeds up Python by about 15%.
Maybe the new thing for Vista is just that the low-fragmentation heap is the default mode.
For applications, heap management is in a library, not the kernel.
Actually, the real problem is not heap management for active programs, it's too much crap idling in the background, chewing up memory.
I/O prioritization?
QNX has had that for a decade. It's essential in a real-time system. I'm sure Microsoft's implemention is far more complicated, though.
American Technology Research, Inc. currently has 60.7% of the securities under research coverage rate as BUY, 37.4% of the securities rated as HOLD, and 1.9% of the securities rates as SELL. -- Coverage list.
Amtech is supposed to have a performance chart page where they show their past ratings and the matching stock prices.
But a sizable fraction of the companies on the list, like Ask Jeeves and AT&T Wireless Services, are no longer separate companies. Generally not for good reasons.
Peter Lynch, of Fidelity Magellan fame, used to have a framed letter on his office wall from a similar firm. It read "We are withdrawing our 'buy' rating because of the bankruptcy."
Visual Basic was supposed to do exactly that. And, to some extent, it did. Many little but important business applications were written in Visual Basic.
Microsoft managed to extend it into a monster, but that's Microsoft.
(Visual Basic, like FrontPage did not come from Microsoft; those were acquisitions.)
It's amusing that the Wikipedia article for "Mashup" begins with "To meet Wikipedia's quality standards, this article or section may require cleanup."
I run Downside, which provides some financial information, mostly about failing and failed companies. For some years, I just routed all the mail for the domain to a default mailbox, regardless of username.
There's a "Downside School and Abbey" in England. I'd get misaddressed e-mail to and from students and teachers. Most of this was minor,
although a few students signed up for things that generated spam,
resulting in my getting junk mail about British teen idols. On one
occasion, though, I got a misaddressed message headed "I am going to kill you tonight". This was after the Columbine massacre, when everyone was paranoid about school shootings, so I called the school, got someone up in the middle of the night their time, and read them the message. It turned out to be a 12 year old kid flaming in E-mail. But you never know.
Then there was "Downside", the band. They had "downside.net" for a while. That produced some amusing E-mail. I'd get some of their mail, even after filtering. I forwarded this to their lead singer, and we'd occasionally exchange messages. That was fine, but then they tried offering branded e-mail to their fans, using the "downside.net" domain.
That led to too much junk, and I talked to them about that.
Finally, they were signed by a label, and changed their name to "Strata" to avoid trademark problems, since I own "Downside" as a registered trademark. There are actually several other bands now calling themselves "Downside", but none of them seem to be very active. (The grunge band on Long Island broke up, another let their domain expire, and the third insists they haven't broken up, but don't play much.)
The worst problem was a "joe job", a company sending out spam to advertise their porno sites. At one point, 16,000 mail bounces per day were coming in, and the mail server was overloading. We put some effort into shutting them down. Since they were paying for ISP accounts with credit card numbers they'd collected from their customers,
ISPs were very cooperative. And owning a registered trademark gives you extra legal leverage.We got them kicked off servers on three continents. We had their domains locked and their DNS turned off.
It took a while to trace them to St. Petersburg, Russia, but after some long international phone calls, all their sites disappeared from the web and were never seen again.
It was not until Google Maps released the satellite data along with their mapping services did the focus shift to more three dimensional mapping.
Er, no. Google bought Keyhole, which has been working for several years now. I've had a paid Keyhole account for years. Google's innovation is giving it away, and offering a more limited version in a standard browser.
You need to ask yourself how many programmers ever got a chance to seriously work with those odd machines.
Very few. (I've used an nCube.)
What we'll probably see on the PS3, at least for early games, is that one SPE is devoted to audio, one is devoted to something like decompressing textures and geometry from the disk, one is devoted to dealing with the network, one is doing the 2D GUI components, and the rest are idle. The PowerPC CPU will be doing all the gameplay and the nVidia chip will be doing the 3D graphics. This is suboptimal, but not hard to implement.
More general distributed processing may come later. It would be nice to be able to do physics in the SPEs, but, as the author of the first physics engine that did ragdolls, I can say that making that work comes under the heading of "not fun".
They're useful in situations where lubrication is difficult or the lubricant has to be sealed away from the environment. Clean rooms, vacuum systems, food and drug processing, cyrogenics, and pressurized gases qualify. They're also useful for large diameter bearings.
These things look and work much like an electric motor. They're not that exotic.
The basic problem with the Cell processor is that the SPEs each have only 256K of private memory, with uncached, although asynchronous, access to main memory. It's the unshared memory that's the problem.
This architecture has been tried before, for supercomputers. Mostly unsuccessful supercomputers you've never heard of, such as the nCube and the BBN Butterfly. There's no hardware problem building such machines; in fact, it's much easier than building an efficient shared-memory machine with properly interlocked caches. But these beasts are tough to program.
The last time around, everybody gave up, mainly because more
vanilla hardware came along and it wasn't worth dealing with wierd
architectures.
The approach works fine if you're doing something that looks like "streaming", such as multi-stream MPEG compression or cell phone processing. If you want to do eight unrelated things on eight processors, you're good.
But applying eight such processors to the same problem is tough. You've got to somehow break the problem into sections which can be pumped into the little CPUs in chunks that don't require access to any data in main memory. The chunks can't be bigger than 50-100K or so, because you have to double buffer (to overlap the transfers to and from main memory with computation) and you have to fit all the code to process the chunk into the same 256K. That's a program architecture problem; the compiler can't help you much there. Your whole program has to be architected around this limitation. That's the not-fun part.
You have to make sure that you do enough work on each chunk to justify pumping it in and out of the Cell processor. It's like cluster programming, although the I/O overhead is much less.
In some ways, C and C++ are ill-suited to this kind of architecture.
There's a basic assumption in C and C++ that all memory is equally accessable, that the way to pass data around is by passing a pointer or reference to it, and that data can be linked to other data. None of that works well on the Cell. You need a language that encourages copying, rather than linking. Although it's not general-purpose,
OpenGL shader language is such a language, with "in" and "out" parameters, no pointers, and no interaction between shader programs.
Note that the Cell processors don't do the rendering in the PS3. Sony gave up on that idea and added a conventional NVidia graphics chip.
(This guaranteed that the early games would work, even if they didn't do much with the Cell engines.)
Since the cell processors didn't have useful access to the frame buffer, that was essential. So, unlike the PS2, the processors with the new architecture aren't doing the rendering.
It's possible to work around all these problems, but development cost, time, and risk all go up. If somebody builds a low-priced 8-core shared memory multiprocessor, the Cell guys are toast. The Cell approach is something you do because you have to, not because you want to.
Slashdot Mirror
IPMI is very powerful. An IPMI session starts with a Presence ping Any machine with IPMI hardware should answer a "presence ping" on UDP port 663. This identifies an IPMI-capable machine, and returns some vendor info. Anyone can send this. This should work even if the machine is "turned off", as long as it has standby power and is on a LAN.
Then, there's a challenge-response authentication sequence. More on this later.
Once you're in, here are some of the things you can do:
There's more. Much more. Basically, you can remotely take over the machine, turn it on, inventory the hardware, load an operating system, boot it up, and talk to it.
IPMI's back channel can do more than this. With some help from the operating system (and yes, it's supported in Windows) you can do more remote administration functions. This is great for administering your data center remotely. But it has darker implications.
Supposedly, most machines are shipped with IPMI mostly turned off, unavailable until a program is run on the machine to load in the keys that enable it. Supposedly.
Thus, all it takes for IPMI to be a "backdoor" is for a set of secret challenge/response keys to be preloaded into the IPMI chip. There's no way to read those keys. Short of taking the chip apart, gate by gate, there's no way to tell if there's a backdoor in there. Or a set of keys might be loaded by the system integrator before shipping the system. You can't tell. So that's where to put a backdoor, where no one can find it.
There's an open source, OpenIPMI, for sending IPMI commands on Sourceforge. Send "Presence pings" to the machines you have and see if they answer.
I'm glad to hear that he's back.
Energy. 80% of electricity in France comes from nuclear plants. Most of the rest is from hydroelectric plants. Cheapest electricity in Europe. France exports electricity. Now that's energy independence.
Back in 1973, at the first "oil shock" of the Arab oil embargo, there was debate in France over what to do. The decision was made to go for energy independence. Unlike in the US, that decision was carried out. And now France is reaping the rewards. They don't have to fight wars for oil.
So in the PS3, the Cell processors aren't doing the rendering. The Cell should render about as well as everything else with a current NVidia part.
Flyovers are easy if you have enough RAM and a GPU. How much RAM did the demo rig have?
What Rackable is really pushing is a system where AC to 48VDC conversion takes place in a unit at the top of the rack, and 48VDC is local to the rack. That, at least, simplifies the cable management.
One big advantage of 120/240VAC power distribution using US standards is that the connectors are standardized and reasonably idiot-proof. That is, if you can plug it in, you won't overload the power cord or the connector, and if you overload the branch circuit, a breaker will trip. Outlet strips have circuit breakers, so you can't overload the cord to the outlet strip without a breaker trip. There are NEMA standard power plugs for 15A, 20A, and 30A circuits, 120/240VAC, and single and three phase configurations. All this is standardized nationally and enforced by the National Electrical Code.
In contrast, there are no simple standards for 48VDC. Most 48VDC gear has big screw terminals. There are no standard plugs and sockets. Somebody, preferably a licensed electrician, has to check all the data plates, add up the current loads, calculate voltage drops, size the wire and breakers, and torque the big screw terminals to the correct torque, using the correct lockwashers. Every time you add or change a load, somebody has to recheck the math. Errors can cause a fire. None of this is all that hard if you have basic power technician skills, but you can't just go casually plugging stuff in.
Although, since the development of the low-cost clamp-around DC ammeter, things have become easier in the DC world.
There are lots of attempts to come up with a new form factor and a market niche to go with it. It's getting wierd. Apple's new iPod speakers have the form factor of a ghetto blaster. (There's a strong resemblance between the Apple Hi-Fi and the 1984 Radio Shack boombox with a dock for a Sony Walkman.)
Myspace will probably get worse to become "competitive".
Social networking should be peer to peer. There's no piracy issue; it's all about ME. No problems with the RIAA or MPAA. There's an open source project for somebody.
Actually, we in the US don't. Crime is down. The aging population is the biggest driver of this.
Residential burglaries are way down because there's no market for used home electronics. Auto theft is down because cars are harder to steal. And some of the big-city police departments, starting with NYC, finally figured out how to manage their cops.
The discouraging thing is that we probably could, today, build automated spacecraft that could reach the asteroid and set off nuclear bombs to change the orbit. It would be profitable to nudge the thing into earth orbit. And if somebody screws up, we lose the planet.
It's not much of a technology. That's called "structured light", and it's been used for years in industrial computer vision systems. It's one of the simplest ways to measure depth in an image.
It's not even new to law enforcement. Here's a PowerPoint presentation on using it to look at stamped logos in pills. This is from a 2004 conference in Dallas.
Google 1, search engine spammers 0.
I divert Bonded Spammer mail to a separate folder. Let's see what's in there:
That's Bonded Spammer's track record. Let's see how AOL does.
Which MMORPG has the biggest world population? That is, which seamless online 3D world has the most people, without "shards" or "instances" to divide up the world?
This may have applications for soft-body physics in games.
Maybe the new thing for Vista is just that the low-fragmentation heap is the default mode.
For applications, heap management is in a library, not the kernel. Actually, the real problem is not heap management for active programs, it's too much crap idling in the background, chewing up memory.
I/O prioritization?
QNX has had that for a decade. It's essential in a real-time system. I'm sure Microsoft's implemention is far more complicated, though.
Amtech is supposed to have a performance chart page where they show their past ratings and the matching stock prices. But a sizable fraction of the companies on the list, like Ask Jeeves and AT&T Wireless Services, are no longer separate companies. Generally not for good reasons.
Peter Lynch, of Fidelity Magellan fame, used to have a framed letter on his office wall from a similar firm. It read "We are withdrawing our 'buy' rating because of the bankruptcy."
Microsoft managed to extend it into a monster, but that's Microsoft. (Visual Basic, like FrontPage did not come from Microsoft; those were acquisitions.)
It's amusing that the Wikipedia article for "Mashup" begins with "To meet Wikipedia's quality standards, this article or section may require cleanup."
There's a "Downside School and Abbey" in England. I'd get misaddressed e-mail to and from students and teachers. Most of this was minor, although a few students signed up for things that generated spam, resulting in my getting junk mail about British teen idols. On one occasion, though, I got a misaddressed message headed "I am going to kill you tonight". This was after the Columbine massacre, when everyone was paranoid about school shootings, so I called the school, got someone up in the middle of the night their time, and read them the message. It turned out to be a 12 year old kid flaming in E-mail. But you never know.
Then there was "Downside", the band. They had "downside.net" for a while. That produced some amusing E-mail. I'd get some of their mail, even after filtering. I forwarded this to their lead singer, and we'd occasionally exchange messages. That was fine, but then they tried offering branded e-mail to their fans, using the "downside.net" domain. That led to too much junk, and I talked to them about that. Finally, they were signed by a label, and changed their name to "Strata" to avoid trademark problems, since I own "Downside" as a registered trademark. There are actually several other bands now calling themselves "Downside", but none of them seem to be very active. (The grunge band on Long Island broke up, another let their domain expire, and the third insists they haven't broken up, but don't play much.)
The worst problem was a "joe job", a company sending out spam to advertise their porno sites. At one point, 16,000 mail bounces per day were coming in, and the mail server was overloading. We put some effort into shutting them down. Since they were paying for ISP accounts with credit card numbers they'd collected from their customers, ISPs were very cooperative. And owning a registered trademark gives you extra legal leverage.We got them kicked off servers on three continents. We had their domains locked and their DNS turned off. It took a while to trace them to St. Petersburg, Russia, but after some long international phone calls, all their sites disappeared from the web and were never seen again.
There's nothing in the rules that says the thing has to run under VMware. So send in something that runs under Xen.
Er, no. Google bought Keyhole, which has been working for several years now. I've had a paid Keyhole account for years. Google's innovation is giving it away, and offering a more limited version in a standard browser.
Very few. (I've used an nCube.)
What we'll probably see on the PS3, at least for early games, is that one SPE is devoted to audio, one is devoted to something like decompressing textures and geometry from the disk, one is devoted to dealing with the network, one is doing the 2D GUI components, and the rest are idle. The PowerPC CPU will be doing all the gameplay and the nVidia chip will be doing the 3D graphics. This is suboptimal, but not hard to implement.
More general distributed processing may come later. It would be nice to be able to do physics in the SPEs, but, as the author of the first physics engine that did ragdolls, I can say that making that work comes under the heading of "not fun".
They're useful in situations where lubrication is difficult or the lubricant has to be sealed away from the environment. Clean rooms, vacuum systems, food and drug processing, cyrogenics, and pressurized gases qualify. They're also useful for large diameter bearings.
These things look and work much like an electric motor. They're not that exotic.
This architecture has been tried before, for supercomputers. Mostly unsuccessful supercomputers you've never heard of, such as the nCube and the BBN Butterfly. There's no hardware problem building such machines; in fact, it's much easier than building an efficient shared-memory machine with properly interlocked caches. But these beasts are tough to program. The last time around, everybody gave up, mainly because more vanilla hardware came along and it wasn't worth dealing with wierd architectures.
The approach works fine if you're doing something that looks like "streaming", such as multi-stream MPEG compression or cell phone processing. If you want to do eight unrelated things on eight processors, you're good.
But applying eight such processors to the same problem is tough. You've got to somehow break the problem into sections which can be pumped into the little CPUs in chunks that don't require access to any data in main memory. The chunks can't be bigger than 50-100K or so, because you have to double buffer (to overlap the transfers to and from main memory with computation) and you have to fit all the code to process the chunk into the same 256K. That's a program architecture problem; the compiler can't help you much there. Your whole program has to be architected around this limitation. That's the not-fun part.
You have to make sure that you do enough work on each chunk to justify pumping it in and out of the Cell processor. It's like cluster programming, although the I/O overhead is much less.
In some ways, C and C++ are ill-suited to this kind of architecture. There's a basic assumption in C and C++ that all memory is equally accessable, that the way to pass data around is by passing a pointer or reference to it, and that data can be linked to other data. None of that works well on the Cell. You need a language that encourages copying, rather than linking. Although it's not general-purpose, OpenGL shader language is such a language, with "in" and "out" parameters, no pointers, and no interaction between shader programs.
Note that the Cell processors don't do the rendering in the PS3. Sony gave up on that idea and added a conventional NVidia graphics chip. (This guaranteed that the early games would work, even if they didn't do much with the Cell engines.) Since the cell processors didn't have useful access to the frame buffer, that was essential. So, unlike the PS2, the processors with the new architecture aren't doing the rendering.
It's possible to work around all these problems, but development cost, time, and risk all go up. If somebody builds a low-priced 8-core shared memory multiprocessor, the Cell guys are toast. The Cell approach is something you do because you have to, not because you want to.
Roland the Plogger strikes again.