Dena Slothower from Stanford University declared that this year's ACM SIGGRAPH 2004 conference, which hosted 27,825 attendees and 229 exhibiting companies, surpassed expectations and boasted a 14 percent growth in attendance over 2003 as well as a nearly 5 percent gain in technical program attendance. "One of the most striking aspects of this year's conference was the incredible diversity of uses for the technical advances in computer graphics and interactive techniques," she noted. Microsoft Research's Hugues Hoppe was honored with SIGGRAPH's 2004 Computer Graphics Achievement Award for his work on surface reconstruction, progressive meshes, geometry texturing, and geometry images, while the University of Washington's Zoran Popovic was given the Significant New Researcher Award for his achievements in computer animation. Recipients of the Outstanding Service Award included Judith R. Brown, formerly with the University of Iowa, and Steve Cunningham of California State University, Stanislaus and the National Science Foundation. The award for Best Animated Short at SIGGRAPH's Computer Animation Festival went to Australian Film, Television, and Radio School student Sejong Park's "Birthday Boy" on the virtue of its outstanding filmmaking and narrative, while Chris Landreth's "Ryan" earned the Jury Award for its use of a new animation and rendering style for storytelling purposes. Science fiction writer and futurist Bruce Sterling delivered a keynote speech at SIGGRAPH 2004 in which he urged the graphics community to re-focus its efforts to bolstering the application of technology to improve the human race. "Our material culture as it is today is not sustainable," he warned.
Click Here to View Full Article

Declining numbers of university computer science enrollments are generating concern that America's technology leadership could be endangered. New fall undergraduate majors in MIT's electrical engineering and computer science department have plummeted from about 385 in 2001 to less than 200 this year, while 2004 enrollments in Carnegie Mellon University's computer science department totaled 2,000--1,200 fewer than three years ago. Peter Lee, associate dean at Carnegie Mellon's school of computer science, fears that computer science doctorate numbers will dip as a result of the erosion of undergraduate computer science majors, which could subsequently cause the United States to lose its edge in computer science research. Furthermore, fewer researchers could reduce emphasis on the discipline of computer science, resulting in less new students. Lee says falling undergraduate enrollment is symptomatic of a general backsliding of student interest in computer science over the last 20 years, which he attributes to the discipline's shift away from speculative research such as machine-enhanced human intelligence and toward more practical, less stimulating, research. The National Science Board recently called attention to a "troubling decline" in the number of Americans pursuing careers in science and engineering, concurrent with increasing numbers of jobs that require skills in those fields. Some observers contend that there is actually a surplus of qualified personnel as the result of layoffs fueled by the dot-com collapse and accelerated offshore outsourcing. However, a recent Rand report found no evidence of shortages since at least as far back as 1990, nor are any expected in the future.
Click Here to View Full Article

The Australian Communications Authority (ACA) is preparing a futures report for the national communications industry that depicts five scenarios set in the year 2020. ACA senior policy advisor Belinda Lester presented the preliminary findings to the ACA Self-Regulation Summit in Sydney, which she said incorporated expert opinions from nearly 200 people located in the United States, Europe, and Asia; she noted that none of the scenarios are likely to play out exactly as forecast, but reality will include aspects from each scenario. The first case presents a communications industry that combines public, private, and individual participation to create completely open and ubiquitous wireless connectivity, while the second scenario is dubbed "Big Daddy" for its dominant government role, where policy is enforced through a few industry players and technology has advanced with breakthroughs such as quantum computing and revolutionized human-machine interfaces. This scenario will likely emerge quickly, and to the surprise of people who did not realize the potential impact of pervasive technology; a new generation that was "chipped at birth" will vie against an older generation that wants to maintain its privacy. In another scenario titled "Nano-boomers," there is a highly pervasive, but not very robust communications environment where information cannot be trusted: This scenario also includes the widespread use of virtual identities so that real and virtual realities are difficult to distinguish. "Marching into the future" is a fourth scenario highlighting nationalistic industry priorities where the search for business applications makes Australia an attractive target for financial and business development, and international talent will likely relocate to Australia. The final scenario is one overshadowed by geopolitical instability, where technology is employed for surveillance and monitoring and large corporations are difficult to rein in.
Click Here to View Full Article

Applying principles of biological evolution to the creation of self-designing machines lies at the core of evolvable hardware, an emergent field whose development stems from advances in computational power and the advent of field programmable gate arrays and other reconfigurable devices. JEM Engineering chief science officer Derek Linden has constructed antennas for NASA's Jet Propulsion Laboratory and military contractors through evolutionary design processes, and he thanks economic considerations for the unexpected interest in evolutionary hardware design: Linden notes that he previously ran antenna models on one Pentium 66 PC, which meant he had to exercise extreme caution with the size of problems and the duration of simulations. "Now, you can brute-force things a lot more easily," he explains. Hardware can be distilled into a simple, binary numeric configuration that can be cross-bred with another design, and the progeny can be tested on a separate software simulator. An optimized design free of shortcomings can be bred after about a thousand generations. Using evolutionary processes, 35 Linux servers designed an antenna at the NASA Ames Research Center in 10 hours with little human assistance, and the antenna will go up in a satellite next year as part of NASA's Space Technology 5 mission. Evolutionary designers put their faith in the computer throughout the entire design process, while traditional engineers only trust the computer for number-crunching and testing. Evolutionary design can also lead to quirky, unexpected results: When University of Sussex researchers Paul Layzell and Jon Bird used genetic algorithms and transistors to create an oscillator circuit, one of the resulting designs diverged from the expected functional profile in that it did not build internal feedback loops to reach the targeted frequency, but rather wired itself so that the hum of a nearby computer was channeled directly through the circuit and into the oscilloscope.
Click Here to View Full Article
(Access to the full article is available to paid subscribers only.)

European researchers have developed communications middleware that allows the integration and ad hoc construction of various mobile services, regardless of the network or hardware platform. The NOMAD project funded by the Information Society Technologies program is due for completion in October 2004, and is currently undergoing usability and infrastructure testing. Project technical manager Ioannis Fikouras gives an example of NOMAD's capabilities: A businessman arriving at the airport is able to find location-based mobile services for both his taxi and hotel room, while the taxi driver automatically receives information about the hotel location and the room is booked remotely. This type of ad hoc action requires automatic service discovery based on location as well as the ability to construct tailored services from simpler ones. In addition to service discovery and compilation, NOMAD also provides a convenient pathway for distributed services, such as data stored in many remote locations being available to any user connected anywhere with a NOMAD-enabled network. The NOMAD researchers are conducting usability and technical tests in Berne, Bremen, and Helsinki, and the team is using NOMAD's distributed services capabilities to link the remote pilots, which make use of a number of sample services. Helsinki Technical University is conducting user testing, and has already uncovered areas for refinement, while the other sites are validating NOMAD for both IPv6 and IPv4 architectures. The focus of the testing is to ensure that NOMAD allows for the best possible user experience while remaining a transparent, supporting technology.
Click Here to View Full Article

Bauhaus University researchers are trying to bring augmented reality technology to the mainstream with a system for mobile phones that interpolates computer-generated 3D models into real-time video on the phone's screen while calibrating the models with physical markers in the camera's field of view. One researcher, Mathias Moehring, envisions such systems acting as museum or metropolitan guides for tourists when IDs encoded into the markers cause the system to display related information about whatever the person is looking at. But cell phone technology must improve before the system can become practical: Moehring notes that cell phone cameras need higher resolution and image quality to ease optical tracking, while their speed and memory must be increased as well. The augmented reality technique aligns 3D graphics within the phone's 2D display through an algorithm that fixes a 3D model to a physical 3D marker captured on video. As the phone moves, the model rotates in parallel with the shifting angle between device and marker, while the model's position is updated at a rate of four to six frames per second. The system displays the specific computer-generated model indicated by a code represented by colored dots contained in the marker; Moehring says the system disregards particular hues and grayscale values and only looks for fluctuations in color intensity that manifest as increases in the video red-green-blue channels. The researchers detailed their research at the Association of Computing Machinery Siggraph 2004 conference.
Click Here to View Full Article

Handsets that integrate cellular and Wi-Fi technology are under development, most notably in Japan, where NTT DoCoMo and Fujitsu are planning to roll out products that bundle the convenience of cellular access with the low cost and high speeds of Wi-Fi. DoCoMo's Hitoshi Yasuda calls the company's new N900iL an "information terminal" with cell phone and PDA functionality that permits users to check office email, schedules, and documents recorded on the corporate server. The phone functions over wireless office networks as well as DoCoMo's high-speed, 3G cellular network. Fujitsu's handset, described as an "all-purpose remote control for the office" by senior researcher Nobutsugu Fujino, can operate on both office Wi-Fi systems and public hotspots, and features a networking card that allows the phone to function on cellular networks when it is outside the range of Wi-Fi transmitters. Fujitsu's Seamlesslink software facilitates a smooth, interruption-free switch between networks; the handset runs on Microsoft's Windows CE operating system, and the device can be customized with different kinds of software. The phone can be outfitted with a card that stores personal ID data, and sensors in the office can pinpoint employees' location by communicating with the device, enabling calls that come in to the employee's work phone number to be routed directly to that person, regardless of location. In addition, when an employee sits down at any workstation, his or her personal computer desktop is automatically raised on the computer monitor. The competition to combine cell phones and Wi-Fi promises to be especially intense in Japan, where cellular technology is highly advanced and almost three-quarters of the populace use mobile phones.

The World Wide Spectrum Efficiency (WWiSE) consortium introduced a proposal for an 802.11n standard Thursday, on the heels of Agere Systems' July disclosure of its own 802.11n proposal. The IEEE 802.11n working group requires that 802.11n throughput surpass 100 Mbps, and the industry expects 802.11n to supplant current Wi-Fi technologies within two to three years. Both the Agere and WWiSE proposals boost throughput by employing a multiple input, multiple output (MIMO) configuration with a minimum of two transmit antennas and two receive antennas that broadcast data in parallel via a 20 MHz channel, and both promise to maintain backward compatibility with current Wi-Fi technology. IEEE will convene in September to start assessing over 60 working proposals for 802.11n's technical infrastructure, and WWiSE members said the first draft of the proposed standard is expected by the middle of next year. WWiSE company Airgo Systems has already started shipping MIMO-enabled chips to its clients, who are planning to ship MIMO-based products. The consortium announced that its members, who include Bermai, Conexant Systems, Texas Instruments, Broadcom, and STMicroelectronics, will retain their own intellectual property (IP), and license it to other members under reasonable terms. There will be no patent amalgamation among WWiSE companies, which means that an 802.11n member will have to accept licenses from other companies. WWiSE members said they intend to "release key messages" about the pre-802.11n standard to other members by publicizing their proposal now. "My suspicion is that there will be several twists and turns as this plays out," said Conexant Systems' Jim Zyren.
Click Here to View Full Article

The U.S. supercomputing industry had become moribund by the time the Earth Simulator from NEC in Japan captured the crown as the world's fastest machine, but Cray, IBM, and Sun Microsystems are competing for a government contract to create a supercomputer that overtakes the Earth Simulator, pulling the U.S. industry out of the doldrums. The three computer makers are semifinalists in the Defense Advanced Research Projects Agency's initiative to develop a petaflop-class computer by the end of the decade, and the agency will fund the construction of real-world prototypes to be ready for use by 2009. A key challenge in inventing a petaflop-class computer is memory lag, a problem rooted in memory technology's failure to keep up with increasing processor speeds. Classic supercomputers that used vector processing--an architecture employed by the Earth Simulator--did not have this problem, but the growing popularity of computer clusters made vector processing passe, thanks to their cost savings as well as their ability to boost computing speed by adding more PCs. Sun expects to overcome memory lag with proximity communications technology, a wireless chip-to-chip methodology that promises to support speeds over 100 times faster than the metal-ball connectors typical of modern chips, as well as reduce heat and power consumption; the trade-off is a greater sensitivity to vibration and slight heat fluctuations, while chips will need near-perfect alignment in order to match up sending and receiving nodes. IBM's planned solution to memory lag is to increase processors' memory cache hundreds of times, although unifying the chips' operation will be a formidable challenge. IBM is also developing self-organizing chips that can function in both vector and cluster configurations. Cray's solution will use true vector architecture, though programming such machines can be difficult.

Zoe, an autonomous, solar-powered rover programmed to look for life in hostile environments, will be demonstrated by Carnegie Mellon University (CMU) robotics and life sciences researchers on Aug. 12 prior to its deployment in Chile's arid Atacama Desert, where it will seek out and identify life forms while traveling 50 kilometers over a two-month period. A team of scientists from CMU, NASA's Ames Research Center, the University of Tennessee, and Chile's Universidad Catolica del Norte will be on hand for Zoe's Chilean mission, which will represent the second phase of a three-year initiative focusing on enabling robots to search for life on Mars. The rover will investigate particularly promising sites in the desert while being remotely guided from an operations center based in Pittsburgh. The guidance team will employ EventScope, a remote experience browser that allows both researchers and the public to perceive the Atacama from the point of view of the rover's sensors and cameras. Zoe is designed to travel at a maximum speed of 100 centimeters per second, which means the machine should cover 2 km per solar day; the robot can also circumvent large obstacles and tolerate a 30-degree incline. Zoe is the successor to Hyperion, another CMU robot whose 2003 deployment in the Atacama yielded results that significantly influenced Zoe's design, instrumentation, and software. Zoe's mission will also contrast life detection skills between robots and people. NASA Ames planetary scientist Nathalie Cabrol explains that the ultimate aim of NASA's Astrobiology Science and Technology Program for Exploring Planets, of which the Atacama initiative is a part, is "to create an astrobiologist without a space suit."
Click Here to View Full Article

Security at the Olympic Games in Greece this month will include street surveillance cameras, paired with sophisticated software, that will act as digital security guards collecting intelligence. The $312 million system was developed by a consortium led by Science Applications International and gathers images and audio from more than 1,000 high-resolution and infrared cameras, four mobile command centers, 12 patrol boats, one blimp, 4,000 vehicles, and nine helicopters. Speech-recognition software will put spoken words into text, and the text and other electronic communications will be searched for patterns. The system covers nine ports, airports, greater Athens, and all the other Olympic cities, and has components used by U.K. and U.S. government intelligence agencies. In preparation for the Olympic Games, the Greek government modified legislation to allow increased tapping of mobile and land line phone conversations. With the technology-enabled security measures and surveillance, authorities will be able to respond to critical incidents in the most effective way since they already have important information on hand, explains Greek police spokesman Col. Lefteris Ikonomou. The camera software is intended to spot and rank possible risks, says Dionysios Dendrinos, general manager of consortium member One Siemens. It is also sophisticated enough to distinguish between a tire blowout and a gunshot. The security net also includes a sensor network established throughout Athens designed to detect chemical agents. There have been some protests over the use of the extended security measures, since some people fear the loss of privacy.
Click Here to View Full Article

Software patent applications continue to take longer to process while the U.S. Patent and Trademark Office works on solutions that would only alleviate pressure slowly over the next few years. Patent industry experts warn that the increasingly lengthy patent process in the United States threatens rather than encourages innovation, and can be especially troublesome to new companies whose business hinges on a single idea. The case of miniFile is a good example, as the company is just now nearing completion of its patent application process after four years; in the meantime, miniFile founder Peter Kostka has found that Plaxo, a startup firm that launched at the same time as miniFile, uses similar technology. If his patent had gone through earlier, Kostka believes the potentially damaging situation between the two companies would have been easier to resolve. Part of the problem is that the U.S. system validates the patent according to when the inventor first conceived of it, not when the actual patent was granted; in Kostka's case, this means having to prove Plaxo's founders came up with their technology after he did. Venture capitalist Kevin Efrusy thinks the patent situation for smaller companies might be overblown, saying, "If a company is relying upon patent protection to save them from competition in the software space, they're in deep trouble." Entrepreneurism and effort are the most important elements to success, he says. A bill pending in the U.S. Senate would allow the Patent Office to keep their collected fees, the majority of which they current dole out to other agencies, and expand operations--however, observers say this would not address the fundamental issue of whether software patents should be granted at all. In Europe, several governments are considering rescinding their votes last month on a European Union law that allows for software patents.
Click Here to View Full Article
(Access to this site is free; however, first-time visitors must register.)

Parham Aarabi of the University of Toronto has developed a technology for masking background noise on cell phone calls so that conversations are clear. The researcher claims his two-microphone system facilitated gains of approximately 30 percent in speech recognition over several other multimicrophone systems in test results, which were detailed in this month's edition of IEEE Transactions on Systems, Man, and Cybernetics. The microphones in Aarabi's system pick up nearby sounds, while processing chips and software choose which sounds should be kept and which should be suppressed; a chip first selects which speaker to focus on (usually the person in closest proximity to the phone), and then a second chip constantly scans all the frequencies through the microphones and determines whether they come from the person holding the phone or some other location. The sounds reaching the mikes are broken into frequency components, and each component's phase is compared, after which frequencies deemed noise are masked while frequencies determined to be emanating from the speaker on the phone are magnified. Microsoft researcher Li Deng, who has devised software algorithms designed to filter out background noise, notes that Aarabi's method is distinctive because it uses phase--the time difference between the signals arriving at the microphones--to a greater extent than any other technique he has seen. Aarabi's technology could also enhance the accuracy of computer and automotive voice recognition interfaces. The University of Toronto researcher says prototype tests demonstrated that the system functions even when the mikes are a mere two inches apart, and he estimates that integrating the functions of mikes, amplifiers, and analog-to-digital converters on single chips would take one to two years.
Click Here to View Full Article
(Articles published within 7 days can be accessed free of charge on this site. After 7 days, a pay-per-article option is available. First-time visitors will need to register.)

The goal of the annual RoboCup competition is to develop autonomous robot soccer teams that can win a tournament, and Cornell University decided this year to embed more intelligence in its robots with the help of an Innovation in Excellence award from Microsoft Research. The additional intelligence will put Cornell's four-wheeled, omnidirectional robots in a better position for future tournaments, in which local vision will be required: As opposed to robots with global vision, which perceive their environment through an camera mounted on top of the field, robots with local vision are equipped with their own cameras. Robots with on-board intelligence are no longer dependent on off-field workstations, and boast improved motion control accuracy, lower system latency, and the ability to probe distributed decision-making strategies. The Cornell machines' increased computing power is provided by a new Windows CE operating system and a PC/104 form-factor computer, and Cornell professor Raffaello D'Andrea explains that the modifications also required the design of a new wireless communication sub-system for debugging. The researchers employed a different drive train for moving the robots around the field, while the incorporation of brushless rather than brushed motors yielded more power density along with a reduction in control over the robots, so drive electronics for the motor had to be designed. The Cornell robots will be enabled for local vision for next year's RoboCup tournament, while their Bluetooth wireless system will be replaced with 802.11a. D'Andrea says that though the robots are exclusively programmed to play soccer, most of the work can be transferred to "real-world" automation through the use of different actuators and sensors.
Click Here to View Full Article

A Zelos Group study published in June draws comparisons between Sun Microsystems' Java programming language and Qualcomm's Binary Runtime Environment for Wireless (BREW), and finds both platforms wanting. Zelos senior analyst and study author Seamus McAteer concludes that BREW's lack of management makes it unviable for vertical enterprise applications, while J2ME's immaturity stems from a shortage of support for classes or components, as well as no architecture for device life cycle management, application management, and data diffusion. "The market will not be defined by limited-function handsets that can accommodate lightweight applications and middleware, but devices that can accommodate full operating systems and will have application processors," McAteer explains, although the study forecasts that BREW or Java will be incorporated into 90 percent of the 100 million mobile devices that are expected to ship over the next five years. Sun's Eric Chu admits that first-generation Java handsets had problems, but claims that MIDP 2.0, due out in the fourth quarter of 2004, supports more consistent behavior and "richer tools." McAteer, meanwhile, argues that Java's development has been severely limited because of its openness, claiming that the Java Community Process has encouraged poor management of J2ME. McAteer notes that the proprietary BREW platform has been much more successful in building markets for billing, data, and content in North America, but its biggest shortcoming is its lack of a clear roadmap for developers--a problem Java does not have. Qualcomm's Jason Kenagy challenges McAteer's assumption that Java and BREW are competitors: He contends that Java is a programming language, while BREW is both supportive of and complementary to Java.
Click Here to View Full Article

Companies are fraught with confusion over the myriad federal regulations they must comply with, and this has spurred business and IT to develop extensible compliance frameworks that can handle any number of regulatory mandates and facilitate reusability that lowers implementation costs and streamlines change management. "Sarbanes-Oxley, the Patriot Act, and HIPAA were the straws that broke the camel's back, and companies are saying, 'We've got to find a better way to do this--the regulations are only going to get worse,'" notes Axentis CEO and Compliance Consortium advisory Chairman Ted Frank. The solution involves the deployment of distinct business and IT infrastructures, with the former being capable of setting up and maintaining internal controls and repeatable processes that guarantee dependable regulatory compliance, and the latter comprising a technology framework that taps existing resources in which point solutions are exceptional rather than standard. Storage hardware comprises the foundation of a modular, extensible IT architecture, and should support document management, email archiving, security, and business process management software and their associated monitoring and change management capabilities. Industry groups recommend that all IT assets be rigorously evaluated from top to bottom before any expenditures are made, as money can be saved by leveraging existing resources. Gartner research director Brian Wood recommends that companies set up the position of chief compliance officer (CCO) whose stature is equal to the CEO, and have the CCO deploy a rotating IT representative to measure existing IT assets, confirm processes, satisfy security requirements, and make sure that abuses can be tackled using clear techniques. Wood's strategy for risk assessment is to assign risk levels to systems, processes, and staff that are vulnerable to intrusions, evaluate potential intrusions' aftereffects, and then organize a priority list for deploying systems and controls.
Click Here to View Full Article

IEEE R&D Policy Committee Chairman and Rochester Institute of Technology professor Ronil Hira and Campus IT Solutions CEO Bob Denis offer differing strategies on how to strengthen the U.S. IT workforce: Hira calls for reforms to guest-worker visa programs, while Denis thinks close collaboration between business and the U.S. educational system is the answer. Hira observes that industry lobbyists and even the National Science Board are ignoring the simple fact that more attractive IT careers--good salaries, interesting work, job security, reasonable schedules, and so on--will ensure a robust IT workforce. Guest-worker visa programs such as TN, L-1, and H-1B have helped stimulate U.S. tech innovation, but Hira argues that supporters disregard the fact that these initiatives are being used to ramp up the practice of offshore outsourcing and supplant domestic labor with foreign labor. He cites a study of the U.S. Department of Labor H-1B database that demonstrates the ineffectiveness of protections such as a provision that requires companies that request guest-worker visas to pay the prevailing wage for the jobs they intend to fill. Hira concludes that "Exporting IT jobs may be inevitable, but we shouldn't let the government accelerate the destruction of U.S. jobs." Denis suggests that industry and education "re-engage" to produce qualified IT change agents, and he offers programs and initiatives to bolster this re-engagement. Among them: The establishment of a curriculum-advisory board that concentrates on real-world subjects such as project management, process engineering, operations management, and systems architectures; allowing students to work on real industry projects; an increase in academic scholarships to encourage more enrollment in IT courses; associate professorships for industry people who can and are willing to serve as teachers; setting up industry boot camps where professors can re-acquaint themselves with changes that significantly affect IT; and emphasizing certain IT majors at particular schools.
Click Here to View Full Article

Conventional Web-searching applications are ill-equipped to accurately store, index, and search files of three-dimensional objects because they can usually only focus on textual components, but university researchers are working on search engines that can sift through databases of 3D objects by fixing on physical characteristics. The key element of 3D object mining is the volume pixel, a set of graphical data that maps out the tiniest building block of a 3D image; computers can only display a 3D object two-dimensionally, and this process requires the division of the object into pixeled 2D cross sections that are stacked according to the appropriate interpixel and interslice distances, after which interslice gaps are filled in through data interpolation. Selected or drawn image queries are translated into mathematical models through algorithms, and these models are then compared to those of catalogued objects. Purdue University's 3D Engineering Search System (3DESS) enables users to query image databases by sketching objects from scratch, modifying only parts of existing objects with a pencil, or choosing an object from a catalog. Once a drawing of the desired object has been furnished, an algorithm extracts key elements to build a reduced graphical representation or skeleton of the object that is stripped down by another algorithm into a skeletal graph, which eases the storage and indexing of descriptions. When 3DESS finds objects in response to a query, users can indicate which objects more closely resemble the desired object, and enable the system to refine current and future searches via neural-network technology. 3D searching is likely to be a popular tool for inventory management, corporate marketing, and intelligence, once technical and marketplace obstacles are overcome. Search accuracy, speed, and ease of use need to be improved, while most individual and corporate computer users do not carry out enough 3D object searches to legitimize investment in 3D search engines.
Click Here to View Full Article

In the 1960s sense of the term, a virtual machine is a software abstraction that resembles a computer system's hardware, whereas the current definition covers a diverse array of abstractions that are all connected by their use of software written to run on the virtual machine. Virtual machines are abstracted by virtualization software that inserts a virtualization layer at various places in the system: Hardware-level virtualization, a layer that is positioned atop the hardware exporting the virtual machine abstraction, ensures that the virtual machine will run all software written for the hardware. Although hardware-level virtualization, which was very popular in both the research and commercial marketplace in the 1960s and 1970s, dwindled away to virtual invisibility in subsequent decades, it has been enjoying a resurgence since the late 1990s in both the server and desktop arenas. The thin layer of software responsible for exporting hardware-level virtual machines is known as the virtual machine monitor (VMM), and the VMM can run virtual machines in a secure, isolated environment thanks to hardware virtualization. Modern VMMs export the primary attributes of software compatibility, which guarantees that the virtual machine will run the real hardware, all operating systems, and hardware-oriented applications; isolation capability, whereby virtual machines are kept isolated from one another via the machine's hardware production mechanism; encapsulation, which enables the VMM to manage hardware resources as well as the whole software stack; and virtual machine performance levels close to those of the real machine. The renewed interest in VMMs can be traced to their support of application compatibility, program testing and development, accelerated application implementation, and logical partitioning, among other things. Past experiences indicate that a balance between deploying functionality with a VMM and deploying it within the operating system should be ascertained, as well as which of those components controls the hardware resources.
Click Here to View Full Article