[bbmf]

Intel's newest platform

Today's launch of the latest version of Intel's vPro platform is a much bigger deal than you might think, with implications for end users that extend far beyond the enterprise arena at which vPro is initially aimed. The 2007 version of vPro represents the culmination of two of Intel's most ambitious and important plans for the PC platform: the transformation of x86 into a fully virtualizable ISA complete with virtualized I/O, and the first fully-complete implementation of all the parts of Intel's controversial contribution to "trusted computing" technology, formerly codenamed "LaGrande" but now called Trusted Execution Technology (TXT).
Let's take a look at the new vPro and what its new virtualization and "trusted computing" capabilities mean for ordinary users.
The hardware
The hardware side of today's launch has gotten the most coverage, but the basic specs and product names are probably the least interesting and important aspects of the announcement. So I'll run through it quickly and refer you to Intel's press release if you want to know more.
Intel has announced three new Core 2 Duo processors to go with the new vPro platform: the the E6550 at 2.33GHz, the E6750 at 2.66GHz, and the 3GHz E6850. These new processors have an idle power consumption rating of 8 watts, which is almost a third of their predecessors' 22W idle power.
The chipset that these processors will be paired with is the new Intel Q35 Express chipset, which features the ICH9-DO southbridge and 82566DM gigabit Ethernet controller. This chipset and Ethernet controller hardware have support for two of vPro's most significant features, which I'll discuss in more detail below.
Virtualization technology for directed I/O
Today's vPro launch marks the introduction of the long-awaited Virtualization Technology for Directed I/O, or VT-d. The VT-d launch is a major milestone for both x86 virtualization and the Trusted Computing Group's vision of so-called "trusted computing." In a nutshell, VT-d is to virtualization what protected memory was to multitasking, and the long-term impact for x86 computing will be just as large.
In the early days of virtualization on x86 there were two major hurdles that any virtualization vendor had to get over. The first of these was the fact that some x86 instructions had effects on the system that could potentially tip off a guest operating that it wasn't running as root. Virtualization vendors like VMware used a trap-and-emulate approach to catch these instructions and prevent a guest OS from learning that there was actually a hypervisor running beneath it. Intel eventually remedied this deficiency of the x86 ISA with the VT-x extensions, and in subsequent iterations of VT-x it has improved the performance of the new VM entry and exit instructions to make them competitive with the older trap-and-emulate approach.

DMA remapping. Source: Intel Technology Journal, vol. 10 iss. 3

The other big problem, which has persisted until now, involves mapping interrupts and DMA transfers to the proper virtual machine and keeping such interrupt and DMA traffic private to each VM. The need for a virtualization-friendly I/O memory management unit (IOMMU) has been significant, and now Intel has addressed that need with Virtualization Technology for Directed I/O, also known as VT-d. VT-d does its thing by controlling access to memory at the physical page level so that I/O devices that are not assigned to a particular VM cannot access that VM's memory space and vice versa. No longer will a device under the control of one VM be able to access another VM's memory space because the IOMMU knows which regions of physical memory correspond to which VM, and it can block or grant access to those regions based on the source of the I/O request. So VT-d fixes the final place where x86 has been hostile to virtualization, making interrupts and DMA transfers fully virtualizable.
Note that the newly announced Ethernet controller works with VT-d, and it also contains some nice features of its own. It has the ability to store security tokens so that a remote manager can connect to and troubleshoot the machine even when the OS, which normally handles the token-based handshake at the other end of a secure network connection, is incapacitated. The new NIC also hosts a few basic traffic filters that let it check for certain types of network activity characteristic of viruses and Trojans.
With x86 now a virtualization-friendly ISA from top to bottom, Intel can slot the last and long-planned piece of its platform remote management/security solution into place. Enter Trusted Execution Technology (TXT), formerly known as LaGrande.
TXT: Because your network admin (and Big Content?) needs a safe place inside your system that you can't tamper with
Given the fact that we at Ars, along with the EFF, Slashdot, and the rest of the usual suspects, have beat the "LaGrande" = "Big Content inside your PC" = "DRM nirvana" drum for going on five years now, I find remarkable the lack of hue and cry over the full-scale, production, mass-market rollout of what was once considered by the geekerati (myself included) to be the Worst Idea Ever. Let's take a brief look back at the history of what was finally announced today.

• Way back in 2002, when Microsoft announced plans to turn your PC into the Panopticon by placing a fortified virtual room—complete with watchers behind a one-way mirror—in your system, an uproar forced the company to scale back its ambitions. The name of the initiative was Palladium, and the idea was that DRM-encumbered content could run in a "trusted," copy-proof chamber and funnel audio-visual output over trusted links to your monitor and speakers.
  
• The Palladium announcement was followed shortly by Intel's announcement of LaGrande, which would be the hardware part of this "trusted computing" vision. The centerpiece of LaGrande is the Trusted Platform Module, a device whose stated purpose is to safely keep secrets from would-be hackers, viruses, and trojans. But the TPM is also designed to keep secrets from you, the end user, because who knows—you may be a filthy pirate who's out to exercise your fair use rights.
  
• In brief, the TPM is a small ASIC that sits on the motherboard and is useful for a number of different security-related functions. In addition to a small pool of nonvolatile storage that can store keys, it has hardware that implements RSA algorithms for key generation, signatures, and encryption/decryption. It also has a small execution engine that can execute the program code for initializing the device and performing the "measurement taking" functions that are essential to its ability to determine if an execution environment is tampered with.
  
• These TPM facilities can be used to boot the machine in a known, "trusted" state, with a "chain of trust" that extends throughout the entire boot process from the BIOS to the hypervisor or operating system. The TPM can stop the boot process immediately if it determines that the code at any link in the chain (BIOS, hypervisor, OS) is not correct, and it can also signal to a remote system if the machine has booted into a trusted state or not.
  
• Earlier versions of the TPM have been in PCs for some time now, but the module is rarely used both because Windows XP doesn't support it and because the TPM alone is of limited utility. With the launch of the new vPro, however, the full force of LaGrande is now officially upon us. The primary missing piece that has now fallen into place for LaGrande to create fully locked-down regions within your computer is support for I/O virtualization, which finally gives Intel platforms the ability to load, run, and display protected code and content in fully sealed, completely trusted (i.e., "trusted" by your network admin, Sony, Disney, Microsoft, etc., to keep you out of the parts of your system where you don't belong) environments now called Measured Launch Environments (MLEs, formerly "vaults" in LaGrange lingo). VT-d protects the memory space of the sealed-off vault, be it a virtual machine or a process hosted by the OS, from access by unauthorized devices, thereby closing the hole that interrupts and DMA transfers had left in previous vPro implementations.

The higher level of control over memory—specifically the ability to control access at the level of individual physical pages—that VT-d gives TXT provides other security features, like the ability to erase all traces of an MLE from memory once it terminates.
You won't really see much mention of DRM in most vPro coverage, because Intel has (quite appropriately) spent a lot of time and effort over the past few years talking up the TXT + VT-x and VT-d combination as a robust enterprise security and remote management solution. And indeed it is. But as a potential technological enabler of more effective DRM, it's also the ideal companion to Blu-ray and HD DVD, and a godsend to Big Content. Look for it across the rest of Intel's desktop and portable line by the end of 2008 and prepare to kiss fair use goodbye.
Everyone who values security will love it
vPro's potential as an enabling platform for intrusive DRM (again, only one of many possible uses) is disturbing enough that it merits much more attention than it seems to have gotten so far in the tech press, but it shouldn't overshadow the platform's potential benefits. Some nice things are already being done with vPro as a basis, like Symantec's Virtual Security Solution (VSS), formerly known as "Project Hood."
The basic idea behind VSS, which was demonstrated at the vPro launch event, is to launch a lightweight hypervisor that runs the same network intrusion prevention software that you find in Norton AV in a secure virtual machine that can't be reached from Windows. The software sits between the NIC and the OS, and filters inbound and outbound network traffic looking for Trojans, keyloggers, viruses and so on. Ideally, VSS would neutralize such malicious software before it gets into Windows, and if it the Windows partition does become infected the virus can't contaminate the antivirus software because it resides in a secure VM.
New applications like VSS, along with potential of TXT and VT-d for secure remote access and management, will eventually make vPro a must-have in the enterprise and on the consumer desktop. The technology opens up whole new vistas of secure network communication that will excite everyone from e-commerce software vendors to corporate IT departments to foreign governments and militaries... er... waitaminute.
According to a report [PDF] by Endpoint Technologies, both the NSA and the U.S. Army now rely on technology from the Trusted Computing Group, with the Army mandating the same TPM v1.2 modules that form the basis of vPro in all of its computers for network security reasons. This being the case, many folks, myself included, look at "trusted computing" technologies and wonder how the federal government allows them to proceed to market absent the export controls that are typical of strong encryption and so-called "dual-use" technologies that could have potential military applications. But that's a topic for another day.
In the meantime, suffice it to say that vPro 2007 is a big deal for enterprise computing in the near- to medium-term, but whether it's a big deal to online shoppers, foreign militaries, or parties on either side of the DRM vs. "piracy" battles is a question that only time will answer.


http://arstechnica.com/articles/paed...ndor-P2P.ars/1

[bbmf]

Model could reduce fabrication costs
Computer chips inside high-speed communication devices have become so small that tiny variations which occur during chip fabrication can make a big difference in performance.
The variations can cause fluctuations in circuit speed and power causing the chips not to meet their original design specifications, says MIT Professor Duane Boning, whose research team is working to predict the variation in circuit performance and maximize the number of chips working within the specifications.
The researchers recently developed a model to characterize the variation in one type of chip. The model could be used to estimate the ability to manufacture a circuit early in the development stages, helping to optimize chip designs and reduce costs.
"We're getting closer and closer to some of the limits on chip size, and variations are increasing in importance," says Boning, a professor of electrical engineering and computer science (EECS) and associate head of the department. "It's becoming much more difficult to reduce variation in the manufacturing process, so we need to be able to deal with variation and compensate for it or correct it in the design."
Boning and EECS graduate student Daihyun Lim's model characterizes variation in radio frequency integrated circuits (RFICs).
RFIC chips are integral to many of today's high-speed communication and imaging devices, such as high-definition TV receivers. Shrinking the size of a chip's transistors to extremely small dimensions (65 nanometers, or billionths of a meter), improves the speed and power consumption of the RFIC chips, but the small size also makes them more sensitive to small and inevitable variations produced during manufacturing.
"The extremely high speeds of these circuits make them very sensitive to both device and interconnect parameters," said Boning, who is also affiliated with MIT's Microsystems Technology Laboratories. "The circuit may still work, but with the nanometer-scale deviations in geometry, capacitance or other material properties of the interconnect, these carefully tuned circuits don't operate together at the speed they're supposed to achieve."
Every step of chip manufacturing can be a source of variation in performance, said Lim. One source that has become more pronounced as chips have shrunk is the length of transistor channels, which are imprinted on chips using lithography.
"Lithography of very small devices has its optical limitation in terms of resolution, so the variation of transistor channel length is inevitable in nano-scale lithography," said Lim.
The researchers' model looks at how variation affects three different properties of circuits-capacitance, resistance and transistor turn-on voltage. Those variations cannot be measured directly, so Lim took an indirect approach: He measured the speed of the chip's circuits under different amounts of applied current and then used a mathematical model to estimate the electrical parameters of the circuits.
The researchers found correlations between some of the variations in each of the three properties, but not in others. For example, when capacitance was high, resistance was low. However, the transistor threshold voltage was nearly independent of the parasitic capacitance and resistance. The different degrees of correlation should be considered in the statistical simulation of the circuit performance during design for more accurate prediction of manufacturing yield, said Lim.
The researchers published their results in two papers in February and June. They also presented a paper on the modeling of variation in integrated circuits at this year's International Symposium on Quality Electronic Design.
The research was funded by the MARCO/DARPA Focus Center Research Program's Interconnect Focus Center and Center for Circuits and Systems Solutions, and by IBM, National Semiconductor and Samsung Electronics.


[SIZE="1"][url]http://www.innovations-report.de/html/berichte/informationstechnologie/bericht-89065.html[/url[/SIZE]]

[bbmf]

SanDisk uSSD 5000 Solid State Drive – An Inexpensive Embedded Module –
Substitutes For Costlier Hard Disks; Will Be Used In The Intel-powered classmate PC

BERLIN, GERMANY, August 31, 2007 – Joining global efforts to bring affordable personal computers to millions of people in developing nations, SanDisk® Corporation (NASDAQ: SNDK) today announced the uSSD™ 5000 solid state drive (SSD), which can be used as an economical substitute for hard disk drives in sub-$250 PCs. Among the first computers using uSSD solid state drives will be the Intel-powered classmate PC, intended for educational markets in emerging nations.
The announcement was made at IFA, the world’s largest consumer electronics show, where SanDisk is exhibiting at Booth 130, in Hall 15.1 of Messe Berlin, through Sept. 5.

SanDisk’s uSSD 5000 solid state drive is a USB module designed to be embedded directly onto the motherboard of low-cost PCs as a hard disk replacement. The uSSD 5000 solid state drive supports a variety of operating systems, including Microsoft’s Windows XP Professional, Windows XP Embedded, Windows Embedded for Point of Service and Windows CE, as well as Linux. The uSSD 5000 solid state drive is expected to be available in capacities from 2 gigabytes (GB)1 to 8 GB.
This is the first SSD product from SanDisk to be offered with SanDisk’s patented multi-level cell (MLC) technology. MLC technology doubles data- storage capacities in the same physical space as single-level cell (SLC) technology, resulting in significant cost savings.
“The low-cost educational PC category is an emerging market for flash storage where low cost, ruggedness and low power consumption will be the primary factors for broad-based adoption,” said Greg Rhine, senior vice president and general manager of the Consumer Products Division at SanDisk. “At 2GB, the uSSD 5000 solid state drive delivers the necessary storage capacity for low-cost PCs at significantly less cost than conventional hard drives, while meeting performance and reliability requirements for this market. We are proud to have been chosen for the Intel-powered classmate PC, a product that demonstrates how inexpensive SSDs can contribute to the worthy cause of improving education in the developing world.”
Compared to conventional hard disk drives, SanDisk’s uSSD 5000 solid state drives are advantageous in low-cost PCs for a variety of reasons:
Cost Effectiveness – uSSD 5000 solid state drives at 2GB are significantly less expensive than even the lowest-capacity conventional hard disk drives. Future generations of flash memory will usher in higher storage capacities at the requisite price points.
Ruggedness – With no moving parts, SanDisk’s uSSD 5000 solid state drives are a perfect fit for kid’s classroom wear-and-tear. The durability of uSSD 5000 solid state drives surpasses mechanical hard disk drives.
Reliability – SanDisk’s uSSD 5000 solid state drives deliver high reliability based on advanced flash management technology including powerful error detection and correction algorithms.
Small – At 27mm x 38mm, uSSD 5000 solid state drives are about one-fourth the size of 1.8-inch hard disk drives.
Strong Performance - uSSD 5000 solid state drives with MLC technology offer performance more than sufficient for the low cost PC market.
SanDisk uSSD 5000 solid state drives are expected to be available for customer sampling in 30 days, with volume availability expected in the fourth quarter, in capacities ranging from 1 to 8 gigabytes. In addition to the uSSD 5000 brand, SanDisk offers other SSD products in 4-, 8-, 16-, 32- and 64-GB capacities. More information about SanDisk SSD products is available online at www.sandisk.com/ssd.

[bbmf]

Burt Rutan to be recognized as engineer of the year

The American Institute of Aeronautics and Astronautics (AIAA) will present five awards for key contributions to space science and technology during the AIAA SPACE 2007 Conference & Exposition to be held Sept. 18 - 20 at Long Beach Convention Center in Long Beach, California.
Burt Rutan, president and CEO of Scaled Composites in Mojave, California, will be recognized as the AIAA Engineer of the Year. The award is presented to a member of AIAA who has made a recent (within the past two years) individual contribution in the application of scientific and mathematical principles leading to a significant accomplishment or event worthy of AIAA's national or international recognition. Rutan was selected for bold innovations and concepts that have provided a low-cost approach for access to space and won the civilian space race with SpaceShipOne.
Antal Bejczy, former technical manager at NASA Jet Propulsion Laboratory in Pasadena, California, will be recognized with the Space Automation and Robotics Award. The award is presented for leadership and technical contributions by individuals and teams in the field of space automation and robotics. Bejzy was selected for establishing the foundation of robotics for space applications and pioneering the development of key enabling techniques to overcome the special challenges of communication delay, operation feedback and unstructured environment.
NASA’s Thermal Protection System Tile Repair and Reinforced Carbon-Carbon Repair Teams, NASA Johnson Space Center, Houston, Texas, will be recognized with the AIAA Space Operations and Support Award. The award recognizes outstanding achievements in the architecture, analysis, design and implementation of space systems. All members of the teams are being collectively recognized for outstanding efforts in development and delivery of TPS repair hardware and techniques providing emergency repair capability of shuttle tile and wing leading edge systems. The award will be accepted by Frank Lin and Brian Mayeaux, project managers.
The XSS-11 Team from the Air Force Research Laboratory, Kirtland AFB, Colorado, and Lockheed Martin Corporation, Denver, Colorado, will be recognized with the AIAA Space Systems Award. The award recognizes outstanding achievements in the architecture, analysis, design and implementation of space systems. All members of the government-led, multi-contractor XSS-11 Team are being collectively recognized for the successful design, development, integration and on-orbit tests of numerous first-time technologies and mission operations techniques supporting critical USAF missions. The award will be accepted by H Vernon Baker, AFRL Space Vehicles Directorate and XSS-11 program manager.
Glen Fountain, program manager, Johns Hopkins University, Applied Physics Laboratory, Laurel, Maryland, will be recognized with the AIAA von Braun Award for Excellence in Space Program Management. The award is presented for outstanding contributions in the management of a significant space or space-related program or project. Fountain was selected for 40 years of outstanding contributions to space flight missions to advance the fundamental understanding of the earth’s environment and solar system.
The awards will be presented at the AIAA SPACE 2007 Conference Award Luncheon on Tuesday, Sept. 19, 2007. Each honoree will receive an engraved medal, a certificate of citation, and a rosette pin.
AIAA SPACE 2007 is co-chaired by The Boeing Company, NASA Jet Propulsion Laboratory, and the U. S. Air Force Space and Missile Systems Center, and supported by The Aerospace Corporation, Raytheon, Space Florida, Stellar Solutions, Inc., Cimarron/DKI, and California Space Authority. Space News is the official conference media sponsor.
AIAA’s premier annual space conference, AIAA SPACE 2007 serves as a venue for senior government and industry leaders and technologists. Under the theme, “Space: The Next 50 Years,” the three-day event will highlight the numerous advances in space technologies and applications over the past half-century, focus on the next 50 years of space exploration, and further refine the vision NASA has started to implement. Leaders from government, industry, and academia will convene to share ideas and offer their perspectives on the political, economic and social issues that must be addressed to take space utlization to the next level.


http://www.eurekalert.org/pub_releas...-tat083007.php

[bbmf]

Die Moto, a motorcycle that runs on biodiesel,
set national and international records on Monday,
cruising to a top speed of 130.6 miles (210+km/h)per hour.

Built by The Crucible, a group of industrial artists in Oakland,
the bike has a modified BMW car engine and a handcrafted
aluminum shell. And it's a green machine, too. Running on
B100, or pure bioediesel fuel, the bike emits 78 percent
less CO2 than a standard diesel engine.
The team eventually hopes to crank it up to 160 miles per hour.



http://popsci.typepad.com/popsci/200...sel-bike-.html

[bbmf]

A laser-based approach could make cochlear implants, which currently use electrical signals, more effective.

About 100,000 profoundly deaf people now hear with cochlear implants, which work by stimulating the auditory nerve with a string of electrodes implanted in the inner ear. While the devices enable many users to converse easily and use telephones, they still fall short of restoring normal hearing. Now scientists at Northwestern University are exploring whether laser-based implants could one day outperform today's electrical version.
The mammalian ear uses neural firing rates as one way of encoding sound. As part of a project funded by the National Institute for Deafness and Other Communication Disorders (NIDCD), Claus-Peter Richter and his colleagues at Northwestern have demonstrated that they can control firing rates in the auditory nerve of animals using infrared laser radiation. They are now trying to establish that it's safe to use for long periods of time and that it can manipulate neural firing rates with enough precision to send useful information to the brain.
With conventional cochlear implants, electrical signals spread in the wet, salty environment of the body, muddying the signal. That makes it difficult to trigger specific populations of nerves inside the cochlea. Further complicating matters, simultaneous pulses in different locations merge with each other, stimulating the cochlea everywhere instead of in the desired locations.

Sound effects: An optical cochlear implant would bypass these inner ear hair cells (above),
which are nonfunctional in many cases of profound deafness. Optical radiation would stimulate
the neurons that are normally triggered by the motion of these microscopic fibers in response
to sound vibrations.
Credit: Steve Gschmeissner / Photo Researchers, Inc.

Engineers work around the problem by triggering only one or two of the 16 or 24 electrodes in the inner ear at a time. It's done so rapidly that the user has the illusion that all of the electrodes are firing, but the result is still a relatively crude simulation of normal hearing. To many cochlear implant users, voices sound mechanical and music sounds washed out.
An infrared laser, on the other hand, can be beamed at nerve fibers with pinpoint accuracy. Furthermore, the directional nature of laser light means that optical pulses in different places won't interfere with each other. The increased precision of neural stimulation would make voices and music sound more natural, and users would be able to converse in noisy environments more easily.
While it's not yet clear why infrared radiation can trigger activity in the auditory nerves, Richter hypothesizes that it heats the cells slightly, opening ion channels in the cell walls and sending an electrical signal down the length of the neuron.
A major question is whether it's safe to stimulate nerves in this way for long periods of time. So far, Richter and his colleagues have shown that auditory nerves in anesthetized gerbils can be stimulated with infrared laser radiation for up to six hours without damage. At present it's not feasible to run the tests for longer, but Richter is planning long-term studies in animals with permanently implanted devices.
The researchers are also figuring out how to precisely control neuron activity with lasers. The ear encodes pitch and loudness not just by firing nerves in particular places, but also by modifying the rate at which they fire. So far, Richter has shown that laser radiation can reliably make neurons fire up to 250 times per second, which is comparable to the rate at which early-model conventional cochlear implants drive neurons.
Human trials are years away, but there are several ways in which infrared technology might be used to build a working cochlear implant. One is to use fiber optics instead of electrodes in an array inserted inside the cochlea, somewhat similarly to the way conventional cochlear implants now use electrodes. Early trials of such a system might involve replacing one or two electrodes of a conventional implant with fiber optics to test their effect. Another is to put an optical fiber bundle in front of the cochlea's round window to stimulate auditory neurons without opening the cochlea. (The round window is a thin membrane in the cochlea that absorbs fluid displacement as sound waves travel through it.)
An even more futuristic possibility is to use gene therapy to make auditory neurons responsive to particular wavelengths of light. At MIT, Ed Boyden has been altering nerve cells' genes so that they fire when exposed to one wavelength of light and stop firing when exposed to another. According to Richter, this approach would require less power to activate cells, which might be safer in the long run. Of course, this approach carries all the caveats that typically accompany gene therapy and would require a way to precisely deliver gene therapy to the relevant auditory cells.
"If proven safe and efficacious, optical stimulation could open up ultra-high density stimulation interfaces for the peripheral nervous system," says Boyden. "The process of combining optics and neurons may also pave the way for many future innovations - moving beyond the ubiquitous electrode to new modalities of neural control."


http://www.technologyreview.com/Biotech/19206/page1/

[bbmf]

Endothelial Cells (Source: Image via Tech.co.uk)
Doctors performing artery bypass surgery replace clogged arteries with veins from the inner thigh of the same patient. This vein, the great saphenous vein, is the longest in the human body and can easily be removed if a patient needs a bypass.
However, this approach has its limitations. The patient must have good collateral blood flow and the very long incision can run nearly the entire length of the leg. Ask a few post operation bypass patients and many of them will tell you that the pain from the removal of the great saphenous is often worse than the heart surgery itself.
Researchers from Tokyo Medical and Dental University and Kanagawa Academy of Science and Technology have some promising research that could make the removal of the great saphenous vein unnecessary in future coronary artery bypass operations.
The technology they are developing allows them to print new blood vessels and capillaries from an ink jet printer. The ink used in the technology is made from artificial cells and medical gel in a solution of calcium chloride. This ink solidifies into a tube with a lining of endothelial cells and an outer case of smooth muscle cells.
The technology isn’t mature enough yet to replace the veins from a patients own body with researchers only being able to create a prototype vessel with an inner diameter of 1 mm and a length of 1 cm. The prototype isn’t strong enough to support blood flow yet either. Tech.co.uk says that once the technology is more mature and robust there is no reason why scientists can’t build up artificial organs from layered, printed sheets using the technology.
The process used in this research is very similar to the research being carried out to help combat diseases like Duchenne Muscular Dystrophy and other autoimmune disease by Carnegie Melon’s Institute for Complex engineering Systems and the Carnegie Mellon Robotics Institute.
This research was aimed at using ink-jet printing technology to print out muscle and bone cells and uses Stem cells as a component of their ink. The printer developed in the Carnegie Mellon research deposited and immobilized growth factors in virtually any design, pattern or concentration. The patterns were laid down on extracellular matrix-coated slides. The slides were topped with muscle-derived stem cells and directed differentiate into different pathways to make bone or muscle.


http://www.dailytech.com/article.aspx?newsid=8653

[bbmf]

Storm worm rivals world's best supercomputers

What good are several million Storm worm infected PCs? According to one researcher, the current computing power of Storm worm's botnet is greater than IBM's Blue Gene supercomputer. "If you calculate pure theoretical throughput," Matt Sergeant, chief antispam technologist with security vendor MessageLabs, "then I'm sure the botnet has more capacity than IBM's Blue Gene. If you sat them down to play chess, the botnet would win."
The Australian publication IT News also quotes Sergeant as saying, "In terms of power, the botnet utterly blows the supercomputers away." He goes on to say that just 2 million of the suspected 50 million Storm worm-infected machines are equivalent to the computing power of the top 500 supercomputers.
In the last few months, antivirus vendors have reported an increase in Storm worm infections. Infected computers are often used to relay spam. They can also be used to attack Web sites in what's called a denial-of-service attack.
More alarming is the amount of control the Storm worm bot-herders apparently have over their creation. "We've seen spikes where the owner is experimenting with something and those spikes are usually five to 10 times what we normally see," Sergeant told IT News. "That means they can turn on the taps whenever they want to."
http://news.com.com/8301-10784_3-977...=2547-1_3-0-20

***

Storm Worm Erupts Into Worst Virus Attack In 2 Years

Storm worm authors are blasting the Internet with two types of attacks, and both are aimed at building up their botnet.

The Storm worm authors are waging a multi-pronged attack and generating the largest virus attack some researchers say they've seen in two years.
"We are basically in the midst of an incredibly large attack," said Adam Swidler, a senior manager with security company Postini. "It's the most sustained attack that we've seen. There's been nine to 10 days straight days of attack at this level."
Swidler said in an interview with InformationWeek that the attack started a little more than a week ago, and Postini since then has recorded 200 million spam e-mails luring users to malicious Web sites. Before this attack, an average day sees about 1 million virus-laden e-mails, according to Postini. Last Thursday, however, the company tracked 42 million Storm-related messages in that day alone. As of Tuesday afternoon, Postini researchers were predicting they would see that day between 4 million and 6 million virus e-mails -- 99% of them associated with the Storm worm.
While the number of spam e-mails has dropped significantly, it's still far above normal levels, so Swidler isn't ready to say the attack is over.
The viruses are not embedded in the e-mails or in attachments. The e-mails, many of them otherwise empty, contain a link to a compromised Web site where machines are infected with a generic downloader. This helps pull the computers into the malware authors' growing botnet, while also leaving them open for further infection at a later date.
"This is designed to add computers to the botnet," said Swidler. "That's first and foremost their goal."
But the Storm worm authors aren't contenting themselves with this one attack vector.
Paul Henry, VP of technologies with Secure Computing, said in an interview that the electronic greeting card spam scam that the Storm worm authors launched early in July is stronger than ever. He noted that a friend of his has a company with 100 users and they're being hit with about 300 e-card spams every day.
"Back in December, we saw a huge spike in e-card spams because of the holiday," he added. "We are at the levels we were seeing back in December right now Most security professionals thought it would show up for Independence Day and then fade immediately, but it's been escalating for the last few weeks. It's definitely a pain point."
Again, the e-card spam message, which install rootkits in the infected computers, are working to build a botnet. Henry could not say if it's the same botnet as the other messages are building.
"I have seen thousands of these e-mails since Independence Day. It's got to be working for them or they wouldn't keep doing it," said Henry.
Just a few weeks ago, the Storm worm authors began trying to trick users with fraudulent e-mails warning unsuspecting users about virus or spyware infections. Users around the world were receiving spam messages claiming that viruses or spyware had been detected on the users' systems. It was another attempt to lure users to malicious sites where their computers could be infected.


http://www.informationweek.com/news/...leID=201200849

[bbmf]

Judge strikes down part of Patriot Act

A U.S. district judge has struck down a part of the anti-terrorism-inspired Patriot Act that requires telephone and Internet service providers to turn over records to the government without telling customers.
Judge Victor Marrero, of the U.S. District Court for the Southern District of New York, ruled Thursday that the Patriot Act provision that allows the U.S. Federal Bureau of Investigation to obtain ISP and telecom subscribers' billing, calling, and Web surfing records without court approval violates the U.S. Constitution.
Marrero ordered the FBI and the U.S. Department of Justice to stop issuing so-called national security letters, or NSLs, requiring ISPs to turn over subscriber records. The NSL program prohibited ISPs from telling customers that they were being investigated.
Marrero delayed the order pending a DOJ appeal of his decision.
The NSL program under the Patriot Act violates the First Amendment to the U.S. Constitution as a restraint on free speech, the judge said in his 107-page order. The program also bypasses judicial oversight of the requests, Marrero said.
The NSL program could allow the FBI to unmask the identity of Internet users posting anonymous comments, obtain all of the e-mail messages of an Internet user, and even find out all the Web sites a user has visited, Marrero said.
"In light of the seriousness of the potential intrusion into the individual's personal affairs and the significant possibility of a chilling effect on speech and association -- particularly of expression that is critical of the government or its policies -- a compelling need exists to ensure that the use of NSLs is subject to the safeguards of public accountability," the judge wrote.
The DOJ is reviewing the decision and considering its options, a spokesman said.
The American Civil Liberties Union, which brought the lawsuit against the DOJ and FBI, praised Marrero's decision. "Courts have a constitutionally mandated role to play when national security policies infringe on First Amendment rights," Jameel Jaffer, director of the ACLU's National Security Project, said in a statement. "A statute that allows the FBI to silence people without meaningful judicial oversight is unconstitutional."
The U.S. Congress passed the Patriot Act less than two months after the terrorist attacks of Sept. 11, 2001. Congress, under pressure from President George Bush, reauthorized the Patriot Act in 2006. Many civil liberties groups have criticized the law for being overly broad and targeting innocent U.S. residents.
Marrero in 2004 ruled that the NSL provisions of the Patriot Act violated the Constitution because they amounted to unreasonable search and seizure. After the 2006 revisions to the Patriot Act, the U.S. Court of Appeals for the Second Circuit sent the case back to Marrero for a ruling on whether the NSL provisions were still unconstitutional.
The FBI's use of the NSL program, which existed before the Patriot Act, has ballooned since the law was passed, Marrero noted. In 2000, the FBI issued about 8,500 NSL requests, and the number of requests rose to 56,000 in 2004, he wrote.
The NSL program's "most troubling" aspect, Marrero wrote, is that it attempts to bypass judicial review of law enforcement requests. The program "reflects an attempt by Congress and the executive to infringe upon the judiciary's designated role under the Constitution," he wrote.


http://news.yahoo.com/s/infoworld/91...uazVvpZ98jtBAF

[bbmf]

Microsoft's anti-piracy spyware
Windows Genuine Advantage
has been shown for what it is ...
a dangerously flawed system which
treats customers with disrespect.

(a rose by any other name carries the same scent)
WGA has never been of any advantage to users and now that it has malfunctioned and returned a massive number of false positives customers have every right to demand that Microsoft end this insulting sham immediately.
There are no doubt a lot of Windows XP users incorrectly tagged as software pirates who are feeling thankful right now that they didn't succumb to the temptation of "upgrading" to Vista. At least they didn't have functionality and features stripped from their system.
As for Vista users, many of them would be quite justified in feeling dudded right now. Assuming the majority of them bought their Vista pre-loaded on a new computer, they must be wondering what exactly does this new operating system give them that they didn't have before - aside from more headaches.
One cannot really talk about WGA without also talking about Vista. With Vista, you get an operating system that requires vastly more processing power and memory to achieve equivalent performance to XP, less availability of drivers for third party peripherals, and, for the extra money it costs, you hand your vendor the right to spy on you and disable your software. Now that's value!
Seriously though, one would think that Microsoft already has enough troubles with irate customers in other areas for the company to continue with WGA. Microsoft is already spending more than $1 billion and counting on shoring up its problems with faulty Xbox 360 consoles.
Although it is yet to be mentioned, there is even a possibility that Microsoft could face a class action suit from Vista users who have been adversely affected by having their legally purchased software disabled. This goes beyond merely inconveniencing users by forcing them to reactivate their software online although that's bad enough. Microsoft's snafu actually stopped some users from being able to do some productive work.
As other commentators have pointed out, the name Windows Genuine Advantage is beyond laughable. It is of zero advantage for users to have Microsoft spying on them. Piracy is Microsoft's problem; it is not supposed to be the problem of legitimate users.
In its own interests Microsoft would do well to consider how much WGA is actually costing the company in terms of good will and weigh that up against how much it is actually doing to reduce piracy. Microsoft may believe that with a massive user base locked into its software it is in an invincible position and can therefore do what it likes. However, the mighty have fallen before and will do so again.
One parting thought. If piracy was such a problem for Microsoft in the days before WGA, how was the company able to grow to the size and position of market dominance it holds today?



http://www.itwire.com/content/view/14178/1023/

[bbmf]

Malicious hackers are producing easy to use tools that automate attacks to cash in on a boom in hi-tech crime.
On sale, say security experts, are everything from individual viruses to comprehensive kits that let budding cyber thieves craft their own attacks.
The top hacking tools are being offered for prices ranging up to £500.
Some of the most expensive tools are sold with 12 months of technical support that ensures they stay armed with the latest vulnerabilities.
Tool time
"They are starting to pop up left and right," said Tim Eades from security company Sana, of the sites offering downloadable hacking tools. "It's the classic verticalisation of a market as it starts to mature."
Malicious hackers had evolved over the last few years, he said, and were now selling the tools they used to use to the growing numbers of fledgling cyber thieves.
Mr Eades said some hacking groups offer boutique virus writing services that produce malicious programs that security software will not spot. Individual malicious programs cost up to £17 (25 euros), he said.
At the top end of the scale, said Mr Eades, were tools like the notorious MPack which costs up to £500.
The regular updates for the software ensure it uses the latest vulnerabilities to help criminals hijack PCs via booby-trapped webpages. It also includes a statistical package that lets owners know how successful their attack has been and where victims are based.
MPack has proved very popular with criminally minded groups and in late June 2007 managed to subvert more than 10,000 websites in one attack that drew on the tool.
Hacking groups also operate volume pricing schemes and discounts for loyal customers, he said.
"It's almost a play-by-play of good business practices of software marketing," he said. "When it comes to the hacking industry and level of business acumen there's no limit to what your money can buy."
Paul Henry, vice president of technology evangelism at Secure Computing, said the numbers of downloadable hacking tools was growing fast.
According to Mr Henry there were more than 68,000 downloadable hacking tools in circulation. The majority were free to use and took some skill to operate but a growing number were offered for sale to those without the technical knowledge to run their own attacks, he said.
But, he added, many hacking groups were offering tools such as Mpack, Shark 2, Nuclear, WebAttacker, and IcePack that made it much easier for unskilled people to get in to the hi-tech crime game.
Mr Henry said the tools were proving useful because so many vulnerabilities were being discovered and were taking so long to be patched.
Little risk
"MPack used more than 12 different vulnerabilities that were launched against any web browser that visited any compromised site," he said.
Many hacking groups were attracted to selling the kits because it meant they took little risk themselves if the malicious software was used to commit crimes.
"The only thing you are going to find is a disclaimer that this was distributed for educational purposes and the user accepts any responsibility for any misuse," he said.
The only risk the hacker groups faced in making the tools available was in having someone else steal them and offer them at a lower price. Already, he said, the sheer number of tools for sale was driving down prices.
Garry Sidaway, a senior consultant at security firm Tricipher, said the success of MPack and the attendant publicity was rumoured to be worrying its creators.
"It was made by a group of friends and they all have regular jobs," he said.
Mr Sidaway said the group would not lose much money if they did stop selling it because they made much more from other lines of business.
In particular, he said, the groups can sell information about unpatched or unknown vulnerabilities in software for thousands of pounds per bug.


http://news.bbc.co.uk/2/hi/technology/6976308.stm

[bbmf]

China is starting to ramp up its scramjet propulsion work—an initiative that will benefit high-speed missile programs while also helping the country to develop advanced aerospace materials, greater computational capabilities and a cadre of young engineers who have matured as a result of cutting-edge engine and aerodynamic challenges.
Building on its ramjet experience, China is embracing the much more difficult task of developing Mach 5 air vehicle concepts in which propulsion and aerodynamics are highly coupled.
As part of this effort, an integrated scramjet model is about to begin testing at up to Mach 5.6 in a new wind tunnel in Beijing.
In addition to the technology and engineering experience to be gained, the mid-term military payoff is likely to be more advanced high-speed tactical and medium-range Chinese missiles, especially for antiship warfare that could threaten U.S. aircraft carriers in the Pacific or operating in support of Taiwan.
“China has the greatest potential to compete militarily with the U.S. and field disruptive military technologies that could, over time, offset traditional U.S. military advantages,” the 2006 Pentagon Quadrennial Review said about overall Chinese military technology initiatives.
And over the next several decades, the scramjet work could eventually provide China with a tactical hypersonic global-strike capability beyond the country’s strategic ballistic missile force. The U.S. has similar goals for its own growing scramjet program.
The Chinese allowed a peek into multiple aspects of their scramjet efforts at the recent American Institute of Aeronautics and Astronautics Joint Propulsion Conference in Cincinnati. Chinese engineers from several research facilities presented about a dozen papers on their scramjet developments, as well as details on the new wind tunnel.
At the same forum, their papers revealed new rocket propulsion research, including work on hybrid systems that use a combination of propellants easier to handle and store than most propellants in wider use today. New insight also was offered on Chinese solid rocket motor technology work, important for both missile and space launch applications.
The Cincinnati meeting differed from a traditional U.S. industry gathering, because nearly a dozen engineers from Iran also submitted papers on Iranian solid and liquid rocket technologies. The Iranian engineers are based at the Sharif University of Technology and the KNT Technical University, both in Tehran. They apparently did not deliver the papers in person. However, as participants, the Iranians have access to all of the highly detailed U.S. aircraft and rocket propulsion presentations made at the conference.
A scramjet (supersonic combustion ramjet) flies at Mach 5 or faster using hydrogen fuel and oxygen out of the air for oxidizer. The engine must combine an advanced ramjet that changes configuration to swallow supersonic flow above about Mach 4.
Advanced ramjet technologies are also important for scramjet development, and the Chinese have been active in this area for decades.
Ironically, one the more interesting historical papers presented at the forum was a detailed description of how the U.S. Air Force and Lockheed combined top-secret ramjet propulsion technologies with segmented solid rocket boosters for the Mach 3 D-21B reconnaissance drones that were launched by modified SR-71s and B-52Hs in the late 1960s (see center photo). The D-21B was specifically developed to gather intelligence over China.
This was the first time details on the segmented rocket booster portion of the D-21B program have been presented publicly, says Robert Geisler of Geisler Industries, who led the analysis with retired Pratt & Whitney and ATK Tactical Propulsion engineers. Segmented boosters use individual circular sections like space shuttle solid rocket motors.
China already has such segmented solid rocket motor and ramjet capabilities today, but scramjets are a much greater challenge.
Although nowhere nearly as advanced as U.S. scramjet work, Chinese activities in this discipline will give the Defense Dept. additional impetus to argue for strong, ongoing U.S. hypersonic propulsion funding. Diverse U.S. technology programs are already underway to support development of the X-51 scramjet test vehicle (AW&ST July 23, p. 23).
As part of the Chinese effort, the engineers say new analytical centers are also being developed. For example, a Hypersonic Propulsion Test Facility has been built to support the scramjet program, according to Xinyu Chang, a senior researcher at the Laboratory of High-Temperature Gas Dynamics in Beijing, where the HPTF is located. Gas Dynamics lab research is specifically oriented to “the development of hypersonic flight vehicles, both aeronautics-and space-related,” according to data from the facility.
Broad studies there are “devoted to the fundamentals of hypersonic and high-temperature gas dynamics including detonation phenomena, supersonic combustion, chemical reactions, shock-wave/vortex interactions and thermal-chemical flow characteristics.” The lab helps lead several Chinese technology programs for scramjet propulsion. This includes basic hypersonic vehicle designs that could mate with a scramjet engine, as well as computational fluid dynamics work to assess the challenge of coupled ramjet/scramjet inlet flow fields at the front of the vehicle.
Scramjet ignition technology and work on cooling the internal walls of a scramjet are also being assessed, the Chinese say. Computer modeling of scramjet combustion instability is also being modeled.
“At the present time, the emphasis on rocket-based combined cycle [RBCC] scramjet research has gradually transferred from research and performance studies to some ground experiments and structures design,” says Wang Houqing, a researcher at Northwestern Polytechnical University in Xian. NPU is one of China’s top aerospace research centers.
“A copper model scramjet is ready for testing” in the new Gas Dynamics Laboratory facility, says Xinyu.
“The facility is to provide high enthalpy [thermal dynamic] model scramjet testing,” he says.
The facility uses a hydrogen/air and oxygen replenishment combustion heater with a flow rate of 3.5 kg./sec., with temperature capabilities up to 2,000K. It can generate test velocities up to Mach 5.6, according to Xinyu.
Many different scramjet combustor configurations have been tested so far, he says. But the new facility will allow complete scramjet engine model configurations to be evaluated instead of just the combustor alone.

Other Chinese scramjet research presented at Cincinnati included:
•Aerodynamic performance of Chinese waverider designs integrated with an inlet. “Simulation studies were conducted to investigate forebody-inlet-isolator performance in an airframe-scramjet integrated hypersonic vehicle,” according to Liu Zhenxia, also at NPU.
•Multicode computational fluid dynamics runs for coupled ramjet/scramjet inlet flowfields. This work models the transition from “ram” to “scram” propulsion. The research is underway at the Beijing University of Aeronautics and Astronautics.
•Research of gas discharge coefficients. This work is being conducted at the College of Aerospace and Materials Engineering at the National University of Defense Technology in Changsha.
•Cross-section design of a controllable hypersonic inlet. The research is being done at the Nanjing University of Aeronautics and Astronautics.
•Scramjet combustion mode translation studies. This work is also part of the scramjet effort at the National University of Defense Technology.
•Hydrogen injection and scramjet ignition testing. The research is being done in the Defense Technology university.
•Thermal and structures studies. NPU is performing heat transfer analysis and overall scramjet thermal structure design, including analysis of different materials used in the scramjet concepts.
•Numerical simulation of combustion instability. This work is also being pursued in Xian.


http://www.aviationweek.com/aw/gener...aw090307p2.xml

[bbmf]

A "sea change in the computing industry"
A new startup out of MIT emerged from stealth mode today to announce that they're shipping a 64-core processor for the embedded market. The company, called Tilera, was founded by Dr. Anat Agarwal, the MIT professor behind the famous and venerable Raw project on which Tilera's first product, the TILE64 processor, is based. Tilera's director of marketing, Bob Doud, told Ars that TILE64 represents a "sea change in the computing industry," and the company's CEO isn't shy about pitching the chip as the "first significant new chip architectural development in a decade." So let's take an initial look at what was announced about TILE64 today, with further information to follow as it becomes available.
TILE64
Tell me if this sounds familiar: a grid of processor "tiles" arranged in a mesh network, where each tile houses a general purpose processor, cache, and a non-blocking router that the tile uses to communicate with the other tiles on the chip. If you've followed my coverage of Intel's Terascale research project—especially the 80-core Polaris prototype—then you know that this description fits what Intel has been working on for the past few years and aggressively publicizing for a year or so.
But the basic tile + processor/cache + router + mesh network idea was pioneered by Dr. Agarwal and MIT's RAW project about a decade ago, and now those same ideas also form the basis for TILE64. TILE64 consists of a mesh network of 64 tiles, with each tile containing a general-purpose processor core and a non-blocking router. The short-pipeline, in-order, three-issue cores implement a MIPS-derived VLIW ISA with a few important and peculiar features.
Tilera's PR department is extremely focused on the mesh network and larger SoC architectures as the initial selling points of the processor, so information on the individual cores is hard to come by. Based on my discussion with Tilera and the diagrams that the company provided (see below), each core has a register file and three functional units: two integer ALUs and a load-store unit. The cores also have a split L1 cache (probably 16K), and a 64K chunk of L2 that has an interesting feature. When there's a miss in one core's L2, the core checks the L2 caches of the other cores for the needed data before propagating the miss out to main memory. In this respect, the L2s collectively act like a large 4MB L3.

As you can probably make out from the diagram above, TILE64 has four DDR2 controllers,
two 10-gigabit Ethernet interfaces, two gigabit Ethernet interfaces, two four-lane PCIe interfaces,
and a flexible I/O interface that can be software-configured to handle a number of protocols.

TILE64 is fabbed on TSMC's trailing-edge 90nm process and runs at speeds from 600MHz to 900Mhz. The launch of a 90nm product at a time when the processor market is moving from 65nm to 45nm was undoubtedly done in order to keep costs down. Tilera won't be able to afford to migrate this product to a smaller process node until they get enough volume to justify the investment.
The initial entries in the TILE64 line are now shipping on PCIe daughterboards for development and production purposes. The processor is also available in lots of 10,000 for $435, and further entries to the TILE family are planned to include different core counts.
Raw roots
It's unfortunate for Tilera that Intel has had such success in publicizing the tile- and mesh-based ideas using Terascale as a branding vehicle, because Agarwal and Co. really did get there earlier. But if TILE64 is true to its Raw roots—and I have some indications that it is—then there are a few more interesting things going on below the surface that are worth looking at.
The basic idea behind Raw, a project that was started well before Moore's Law stopped delivering huge clockspeed increases, was that as the number of transistors on a chip increases, wire delay becomes an architecturally significant factor in chip design. The Alpha 21164, Pentium 4, and PowerPC 970 are all examples of the first wave of commodity processors that had made major microarchitectural concessions (i.e., dedicated pipeline stages and increased load-use latencies for certain sequences of integer operations) for wire delay, but the effects of wire delay were still hidden from programmers as far as possible.
Agarwal's idea was to expose wire delay to programmers via the ISA. The Raw project, the name of which seems to be a recursive acronym for "Raw Architecture Workstation," exposes wire delay to the programmer as hops on an on-chip mesh network. It takes one cycle for data to move from one tile to the next, with the result that a compiler can statically schedule operations among multiple tiles' ALUs by taking into account the exact number of cycles that it takes for a result to propagate across the chip.
TILE64 inherits this one-cycle-per-tile feature of Raw, so this type of static scheduling is still possible if you want to write to the bare metal. However, none of this is publicized in the Raw press materials, with ease of programming via Linux and an ANSI C toolchain being emphasized instead.
Note that Raw also had a special bypass network that could take a result from one tile's ALU and route it directly into another tile's ALU so that the compiler could use this network to schedule dependent integer ops. This is a neat trick, and the TILE line inherits this ability; there are four special registers to which an ALU can write to have data sent out directly over the network to another tile.
Wire delay isn't the only microarchitectural feature that Raw exposes to the compiler/programmer. Indeed, Raw was a pretty thorough attempt to repackage the RISC philosophy of "show everything to the compiler and let compiler writers manage the complexity" for the +100-million-transistor era. Even the pin-outs in the package were put under software control as "ports." The result was that you could use Raw to create a kind of "software ASIC" by compiling what Raw programmers referred to as a "software circuit"—an application that's totally tuned down to the cycle level to fit a specific Raw implementation.
I'm due to talk to the head of Tilera's software team, which is actually larger than the company's hardware team, later today, so I can post an update when I find out more. But my sense is that all of this software complexity is still there in TILE64, but it's hidden from most programmers by a very complex and carefully written toolchain that keeps Raw's "network hops" and "ports" and memory hierarchy management as far away from most programmers as possible.
Performance and market positioning
TILE64 is initially being pitched at the embedded market, with wire-speed network processing and HD media encoding being the two main application scenarios that Tilera wants to see it used in. Each TILE64 processor is capable of encoding two simultaneous streams of H.264 video, and over ten streams of broadcast-quality high definition video.
Tilera claims that TILE64 shows a 30x performance per watt advantage over a 3GHz Xeon running a SNORT benchmark, with the new chip able to run the benchmark at 10Gbps speeds. (Note that I've asked for more details on this benchmark run, so I'll publish them when the company gets back to me.) Also claimed for TILE64 is a 40x performance advantage over a TI DM648 DSP chip on a 16x16 SAD (sum of absolute differences) benchmark.
I have to confess that Tilera's choice of benchmark bake-off opponents is a bit odd to me, but now that the product is shipping, I expect to see it benched against, say GPGPU products or IBM's Cell in media encoding bake-offs and chips like Sun's UtraSPARC T2 in massively multithreaded integer and floating-point workloads. After all, the US T2 can handle 64 simultaneous threads of execution just like the TILE64, and it contains similar network and memory interface hardware, so the two should be a good match-up. Tilera claims that TILE64 dissipates between 170 and 300 milliwatts per tile, which compares very favorable to the US T2's already low 2 watts per thread.
Conclusions
Every many-core processor story that I or anyone else writes nowadays is really a software story; efficiently and effectively programming many-core chips is by no means a solved problem, and my money says that the Raw "expose everything to software" approach makes the challenge that much greater. So what will make or break Tilera is not how many peak theoretical operations per second it's capable of (Tilera claims 192 billion 32-bit ops/sec), nor how energy-efficient its mesh network is, but how easy it is for programmers to extract performance from the device. That's the critical piece of TILE64's launch story that's missing right now, and it's what I'll keep an eye out for as I watch this product make its way in the market.
Though there are any number of questions about this product that remain to be answered, one thing is for certain: TILE64 has indeed brought us into the era of 64 general-purpose, mesh-networked processor cores on a single chip, and that's a major milestone. So take a good look at TILE64, because regardless of what happens to Tilera, this is probably what the many-core era looks like.
Update: I originally speculated that the processors cores each had two ALUs and an FPU, but that's incorrect. A Tilera rep has informed me that each core has two ALUs. I've also briefly updated the relevant parts of the article with some new, post-launch information about the processor, and I look forward to posting a follow-up that goes into a little more detail on the microarchitectural and software aspects of TILE.


http://arstechnica.com/articles/paed...4-core-CPU.ars

[bbmf]

Which ISPs Are Spying on You?

The few souls that attempt to read and understand website privacy policies know they are almost universally unintelligible and shot through with clever loopholes. But one of the most important policies to know is your internet service provider's -- the company that ferries all your traffic to and from the internet, from search queries to BitTorrent uploads, flirty IMs to porn.
Wired News, with help from some readers, attempted to get real answers from the largest United States-based ISPs about what information they gather on their customers' use of the internet, and how long they retain records like IP addresses, e-mail and real-time browsing activity. Most importantly, we asked what they require from law-enforcement agencies before coughing up the data, and whether they sell your data to marketers.
Only four of the eight largest ISPs responded to the 10-question survey, despite being contacted repeatedly over the course of two months. Some ISPs wouldn't talk to us, but gave answers to customers responding to a call for reader help on Wired's Threat Level blog.
Marc Rotenberg, the executive director of the Electronic Privacy Information Center, says ISPs should be more circumspect about keeping user data. Maintaining detailed data for long periods of time makes any internet company a huge target for law enforcement fishing expeditions.
"From a user perspective, the best practice would be for ISPs to delete data as soon as possible," Rotenberg said. "(The government) will treat ISPs as one-stop shops for subpoenas unless there is a solid policy on data destruction," Rotenberg said.
The results:
AOL, AT&T, Cox and Qwest all responded to the survey, with a mix of timeliness and transparency.
But only Cox answered the question, "How long do you retain records of the IP addresses assigned to customers."
These records can be used to trace an internet posting, website visit or an e-mail back to an ISP's customers. The records are useful to police tracking down child-porn providers, and music-industry groups use them to sue file sharers. Companies have also used the records to track down anonymous posters who write unflattering comments in stock-trading boards.
Cox's answer: six months. AOL says "limited period of time," while AT&T says it varies across its internet-access offerings but that the time limits are all "within industry standards."
Comcast, EarthLink, Verizon and Time Warner didn't respond.
Some of the most sensitive information sent across an ISP's network are the URLs of the websites that people visit. This so-called clickstream data includes every URL a customer visits, including URLs from search engines, which generally include the search term.
AOL, AT&T and Cox all say they don't store these URLs at all, while Qwest dodged the question. Comcast, EarthLink, Verizon and Time Warner didn't respond.
When asked if they allow marketers to see anonymized or partially-anonymized clickstream data, AOL, AT&T and Cox said they did not, while Qwest gave a muddled answer and declined to answer a follow-up question. Comcast, EarthLink, Verizon and Time Warner didn't respond.
This question was prompted by hints at a web-data conference last March that ISPs were peddling their customer's anonymized clickstream data to web marketers. Anonymization of data such as URLs and search histories is not, however, a perfect science. This became clear last summer when AOL employees attempted to provide the search-research community with a large body of queries that researchers could mine to improve search algorithms. AOL researchers replaced IP addresses with different unique numbers, but news organizations quickly were able to find individuals based on the content of their queries.
Wired News also asked the companies if they have been in contact or discussions with the government about how long they should be keeping data. The Justice Department, along with some members of Congress, are pushing for European Union-style data-retention rules that would require ISPs to store customer information for months or years -- a measure law enforcement says is necessary to prosecute computer crimes, such as trading in child pornography.
ISPs were nearly universally reluctant to talk about any conversations or meetings they have had with federal officials. AOL had no comment, Qwest dodged the question, AT&T wouldn't say, but noted it would broach the issue with the government as part of an industry-wide discussion. For its part, Cox says it has not been contacted.
As for whether they oppose data retention: Qwest said that the market should decide how long data is kept, while Cox was "studying the issue"; AOL is working with the industry and Congress, and AT&T is "ready to work with all parties."
Internet surveillance recently got easier, as the deadline passed last week for ISPs to equip their networks to federal specifications for real-time surveillance of a target's e-mails, VOIP calls and internet usage -- as well as data like IP address assignment and web URLs. While law enforcement currently prefers to ask for stored internet records rather than get real-time surveillance, that balance may shift once the nation's networks are wired to government surveillance standards.


http://www.wired.com/politics/online...05/isp_privacy

[bbmf]

These days, nothing worries an internet service provider more than peer-to-peer file trading.
Depending on where you live, P2P can account for between 50 and 75% of broadband internet traffic. We mostly have the popularity of BitTorrent to thank for this crazy amount of data going to and fro.
This amount of traffic can raise the ISPs daily costs of delivering service, cause congestion either in your neighborhood or on the ISP's network, and force the ISP to buy increased bandwidth capacity.
But if you've been paying close attention to your BitTorrent transfers lately (or if you've simply been reading the news) you'll notice that ISPs have begun to take drastic measures to slow that flood of data currently clogging up their pipes.
Even though many of them deny it, most ISPs actively engage in traffic shaping, bandwidth throttling, connection denial or some such tactic to keep the amount of bandwidth consumed by high traffic applications on their networks to a minimum. While this does often ensure better performance for everyone in the neighborhood, it can mean painfully slow transfer speeds for those dabbling in P2P -- legit or not.

While there are valid arguments for and against shaping, we're not here to debate. We just want the fastest BitTorrent transfers possible.

Methods of Fooling ISPs
So how to get around an ISP that's throttling your BitTorrent traffic? You can try encrypting or your traffic, changing the default port number, changing the way the protocol behaves, reducing the amount of one-way traffic, or hiding your traffic within an encrypted tunnel.
Of course, different ISPs are employing different methods of control. None of these methods are guaranteed to work. But each one is known to work for some, and they are certainly worth a try.
How To Encrypt to Your BitTorrent Transfers:
The RC4 encryption offered by many popular BitTorrent clients today will obfuscate not only the header but the entire stream, which makes it considerably more difficult for an ISP to detect that you're using BitTorrent. Even if your ISP does not force you to enable encryption, you may be connecting to peers with ISPs that do.
Encryption began appearing on clients in late 2005. By the end of 2006, most actively-developed clients were updated with encryption. While not all torrent clients in a swarm will support encryption, most of them will. As a result, this small percentage of non-encryption capable peers may be a reason not to force encryption on a full-time basis, but there is no reason not to enable encryption that allows the falling back to a non-encrypted connection when needed.
If your favorite client is not listed below, check your documentation.
Azureus/Vuze
Azureus (which now calls its official client Vuze) is written in Java and therefore cross-platform. To turn on encryption, head to the Tools menu. Select Options, then Connection, then Transport Encryption. Check the "Require encrypted transport" box and select RC4 in the "Minimum encryption" drop-down menu.
Azureus/Vuze also offers an "Allow non-encrypted outgoing connections if encrypted connection attempt fails" option, which means you'll still be able to hop on torrents that don't have any encrypted seeders.
µTorrent
µTorrent (and now BitTorrent which is based on µTorrent) is a Windows-only client. In µTorrent, open up the Preferences panel and select the BitTorrent tab. Select Protocol encryption and then choose between "enabled" and "forced." µTorrent's "Enabled" option mirrors Azureus' option to allow unencrypted connections when no encrypted clients exist. It will give you more connections, but it won't be as effective at defeating traffic shapers.
µTorrent/BitTorrent also offers a option to 'Allow legacy incoming connections' which lets non-encrypted clients connect to you. This improves compatibility between clients but again, makes your traffic more vulnerable to shapers.
BitComet
BitComet is another popular Windows Client (98/Me/2000/XP). To turn on encryption in BitComet, head to the Options menu and choose Preferences. Then go to Advanced > Connection and select "Protocol encryption." There are options for "auto detect" and "always."
As with the others, "auto detect" will connect to more peers, but it won't hide traffic as well. You'll need to play with the settings in your program to see if it has any affect on your download/upload speeds.
Other clients that support encryption include KTorrent (Linux), rTorrent (Linux, Mac) and BitTornado (Windows).
How To Change Your BitTorrent Port Number:
The default port for BitTorrent transfers is port 6881, with some clients using different ports within the range of 6881-6999. As a result of ISP interference, all clients allow you to change the port number (or port range, sometimes) used for BitTorrent transfers. The setting is in the Options or Preferences for your client, or can be set using a command-line parameter.
Whenever you change your port, you need to adjust your router to allow incoming connections. An excellent service at http://www.portforward.com/ can guide you through the entire process of locating the current port being used (which allows you to change it), and then configuring your router to match.
How To Change the Way the BitTorrent Protocol Behaves:
The BitTorrent protocol has a distinct handshake. To control uploading by seeders, ISPs have learned to look for this handshake. The recent releases of both µTorrent and Azureus/Vuze include a "Lazy Bitfield" feature to hide seeders from ISPs. When Lazy Bitfield is enabled, the handshake is changed to make a BitTorrent seeder initially appear to be a non-seeding peer (sometimes called a leecher). This is done by sending a bitfield indicating missing pieces. Then, once the handshake is done, the client notifies its peer that it now has the pieces that were originally indicated as missing.
Azureus/Vuze
Azureus (which now calls its official client Vuze) is written in Java and therefore cross-platform. To turn on encryption, head to the Tools menu. Select Options, then Transfer. Enable Lazy-Bitfield here.
µTorrent
Lazy Bitfield is controlled in the Advanced section Preferences: peer.lazy_bitfield.
How To Reduce the Amount of One-Way transfers:
Most downloaders become seeders when they have 100% of the archive, then they spend the next several hours "paying back" the swarm until they have provided at least as many bytes uploaded that they downloaded -- a ratio of 1:1 or 1.00. As mentioned before, some ISPs make efforts to control seeders. Seeders generate one-way (outbound) traffic, and this traffic is sometimes the most troublesome for ISPs to handle.
Most clients are configured with a "speed limit" set Upload Maximum Limit in kB/s and an unlimited Download Maximum Limit. To reduce the amount of one-way transfers, the client needs to upload at the same rate (or less, overall) than it is downloading. While this means that the download will be a lot slower to complete, it also means that it will complete at a 1.00 ratio or above.
For example, perform your transfer with an Upload Limit of 30 KB/s and a Download Limit of 25 KB/s. When you first join you won't upload at all because you have no pieces to share yet. But after several minutes, the total bytes uploaded should be equal to or above the total bytes downloaded. When your download is complete, you will have little or no obligation to continue seeding as you already have uploaded enough to the swarm.
This tactic is not always effective or efficient. Some swarms have too few peers left that need data, making it difficult to reach your desired upload rates.
Many multi-torrent clients (Azureus/Vuze, µTorrent, BitComet, and others) provide the option of setting maximum upload and download rates on a per-torrent basis. These settings are found either in a right-click menu or in the Properties of each torrent. Some clients also allow Global Settings that affect all torrents being managed by the client, however the Global Settings do not provide a correct balance to ensure that a one-way transfer is avoided.
Azureus/Vuze provides the additional useful option of limiting the number of seed connections while downloading. This setting is found on the Options panel of each individual torrent.
How To Hide BitTorrent within an Encrypted Tunnel:
With the advent of Application-Layer Inspection, some ISPs may recognize and control BitTorrent traffic despite your best efforts.
You may be able to hide the BitTorrent traffic in an encrypted tunnel -- a transport path within the normal transport paths provided by TCP and IP. You can tunnel your traffic through cooperatives such as The Onion Router (TOR)* or I2P. Commercial Virtual Private Network (VPN) providers such as Relakks or SecureIX will also help keep your ISP from detecting exactly what you're doing. If you are familiar with SSH and SSH Tunneling, this is also a possibility. However, some ISPs even throttle or inhibit these encrypted tunnels.
Azureus provides in-client support for TOR and I2P. Other clients will have to set up the software as recommend on the TOR or I2P site.
*Note: TOR has been updated to allow peer-to-peer download data, despite any information to the contrary (it used to be prohibited).
Now For the Bad News
ISPs are taking advantage of more sophisticated shaping technology all the time, and many of the newer shapers won't be fooled by encrypted traffic. For instance, Sandvine (the shaping tool many believe Comcast and other ISPs employ) won't be fooled by obfuscating your traffic.
So what can you do beyond obfuscating?
The short answer is not much. There is no fool-proof way to do beat the shapers. You best choice is probably to switch to an ISP that doesn't employ anti-BitTorrent traffic shaping. In the long run, this also has the benefit of sending an effective message to your ex-ISP.
For a list of ISPs to avoid, have a look at the list maintained on the Azureus Wiki.
But what if you have no other option when it comes to ISPs?
Start by calling customer service. Call now and call often. Disgruntled consumers often cost an ISP far more money than a large amount BitTorrent traffic. And by all means, try these ideas, though your results may vary considerably depending on what shaping tools your ISP is employing.


http://howto.wired.com/wiredhowtos/i...;category=Play