Pages: pp. 20-22
Researchers have developed a new digital-content protection technology that would solve a key problem with some of the existing approaches, which are designed to eliminate the unauthorized usage, copying, and distribution of multimedia and text.
A team at the University of Maryland's A. James Clark School of Engineering developed the approach, which embeds a unique identifier, or fingerprint code, in a digital image, video, audio, or document file. The forensic code is designed to identify authorized recipients of material and thereby let content owners trace the users who have utilized the file in unauthorized ways.
Figure A new technique helps fingerprint-based digital-content-protection technology avoid collusion attacks. A content owner starts by embedding a fingerprint code, which identifies a specific authorized user, into a file that is subsequently distributed. The owner can thus identify users who utilize files in unauthorized ways. In a collusion attack, however, multiple users take pieces of their files that don't include the entire fingerprints and reassemble them into a new file they can misuse without fear of tracing. The new technique would embed multiple bits of a fingerprint code through overlapping parts of a multimedia file, making collusion attacks more difficult.
The University of Maryland research is designed to thwart collusion attacks, in which multiple users conspire to electronically steal and distribute copyrighted or other sensitive material protected by fingerprints, explained professor K.J. Ray Liu. Colluders obtain copies of protected material, such as a movie, and assemble pieces of each copy. Most or many of the pieces do not have the entire embedded digital fingerprint, which dilutes the protection.
In other projects researching ways to resist collusion attacks, the fingerprint code is too long to easily embed into parts of a multimedia file, explained University of Maryland assistant professor Min Wu.
The University of Maryland technique uses shorter IP-protection code and turns it into a small spreading signal, which can embed multiple bits of a fingerprint code through overlapping parts of a multimedia file. This makes it difficult for colluders to piece together file parts that don't have the full fingerprint code.
Scientists say the new technique could also be used to investigate leaks of sensitive corporate and government data. The US Department of Defense has helped fund the research.
"The commercial potential of any technology that could provide measurable benefits in protecting digital intellectual property would be significant," said Ray Wagner, research vice president for information security and privacy with Gartner Inc., a market research firm.
However, he added, "I have yet to see a forensic method that was not attackable."
Nonetheless, said Wu, the University of Maryland approach appears to be a reliable and effective way to provide some protection against collusion attacks, particularly compared to prior or alternative approaches.
Several companies have reportedly expressed interest in the technology, which is undergoing further research. Liu declined to discuss this.
The US Department of Energy has announced a project that will utilize computer technology to monitor participating consumers' electricity consumption in real time and enable them to decrease usage, particularly during periods of strain on power systems.
As part of the DoE's Pacific Northwest GridWise Demonstration project, 300 homes in Washington state and Oregon will test new intelligent technologies for working with the electrical grid. This approach uses information technology, such as network communications and control applications, to change, economize, and automate consumer electricity use.
The projects' key is enabling home infrastructures such as heating systems, as well as major appliances such as clothes dryers, to recognize and automatically respond to changes in system supply and demand, explained David Chassin, staff scientist with the DoE's Pacific Northwest National Lab. Users could also manually program their electrical systems to conserve energy.
The DoE effort will require the use of specialized equipment. For example, for a participating home, a gateway will provide Internet connectivity to wireless electric meters and the project's back-end servers that handle supply, demand, and pricing information.
The project also includes the Grid Friendly Appliance program. Participating volunteers receive a computerized Whirlpool dryer, a version of which is already on the market. The dryer has a chip that controls the appliance's operations and that includes a sensor that recognizes changes on the electrical line that indicate power-grid instability.
When such instability occurs, the chip saves energy by turning off the appliance's heating element while continuing to tumble the clothes, said Gale Horst, engineering lead in Whirlpool's Research and Engineering Department.
The new project is important because consumption throughout regional power grids can change quickly. Severe grid instability occurs a few times a year, while smaller fluctuations take place every day, Chassin noted. Thus, enabling consumers to adjust their power usage could avoid blackouts and keep utility companies from having to build new facilities, whose costs would be borne by customers.
The program would also help consumers by letting them reduce usage when the load on a power grid, and thus the resulting utility rates, are highest.
The project will study a mix of energy-reduction incentives—such as those based on either lowering overall power usage or responding to electricity supply-and-demand changes—to determine which ones generate the best consumer responses. The DoE—which will spend $1.5 million on the one-year project—plans to pay participants an average of $150 to reduce energy use during peak periods.
Whirlpool, IBM, and Invensys Control have donated equipment, software, and services to the project.
Spectrum allocation is an issue of increasing concern as the number of radio-based devices—including cell phones and computers with wireless Internet connections—proliferate while the limited amount of available frequency stays the same.
University of California, Santa Barbara, associate professor Haitao Zheng is working on a solution to this problem: software-based cognitive radios.
Rather than operate in a preset frequency range, which might experience heavy traffic and interference, cognitive radios use software to dynamically detect and exploit unused or underused radio frequencies. This technique lets the radio select the frequency that will best provide stable bandwidth with minimal interference for the task at hand, Zheng explained.
"Everybody is experiencing this problem because the spectrum is limited," said Zheng. As a result, for example, large numbers of users in the same location—such as a coffee shop—frequently have trouble gaining Internet access via available Wi-Fi connections.
"Even if you assign spectrum," Zheng explained, "utilization is very low. More than 70 percent of the spectrum is not used. The problem is not scarcity. It is that the spectrum is underutilized and overallocated."
She has created algorithms that let disparate devices recognize existing local wireless activity and find unused or underused frequencies. They also let devices negotiate with other radios for the optimal use of spectrum ranges that will minimize interference problems, all without human intervention. In addition, the algorithms recognize noncognitive radios that don't have the intelligence to change their spectrum usage.
In the process, one cognitive radio might exchange the frequency range it's using for a different range if another device requires it. In a particularly busy situation, all devices might give up some normally used frequency to accommodate new participants.
Any radio-based device could use the approach.
Cognitive radios can work with various wireless standards. And Zheng expressed hope that cognitive-radio technology will have its own standards eventually.
She said cognitive radios might not appear commercially for five to eight years. The major obstacles are developing both low-cost radios capable of spectrum sensing and robust protocols to protect against malicious attacks and other types of misuse.
In 1994, historian Ralph Erskine issued a public challenge in a cryptographic journal: crack three yet-to-be decoded messages sent from World War II German submarines. Twelve years later, the M4 Project has used distributed-computing techniques to decode two of the three messages.
The M4 Project takes its name from the M4 four-rotor model of the electromechanical Enigma portable encryption encoder/decoder that the Germans used during World War II. Submarines generally used the M4 Enigma to transmit their coordinates or enemy ship positions.
The UK had an entire unit devoted to breaking transmissions from Enigma machines, which encoded messages via a complex letter-substitution scheme. However, the M4 Enigma, introduced in February 1942, offered such strong encryption, it took the British until December 1942 to start decoding its messages, noted Stefan Krah, a German living in the Netherlands who originated the M4 Project. During that 10-month period, the Germans transmitted the three signals on which Krah's project worked.
To cope with the strong encryption, the M4 Project used two decoding techniques. The brute-force technique generates and tries using every possible decryption key. Hill-climbing algorithms explore various potential solution paths in a step-by-step effort to solve a problem.
Krah wrote all the necessary software and is administering the program. The M4 Project distributed its code-breaking efforts over a set of participating computers, each of which downloaded a client that ran the software.
The project decoded one message ( www.bytereef.org/ m4-project-first-break.html) in 42 days by using 45 participating machines' computing resources, according to Krah. A single 1.2-GHz computer would have required 2,907 hours to complete the task.
M4 subsequently broke the second message ( www.bytereef.org/ m4-project-second-break.html) in four days by using about 2,500 computers' resources.
The third of Erskine's challenge messages has yet to be decoded. The M4 Project might also work on other decryption challenges, including 22 unbroken World War II German Army messages.
Krah said he welcomes volunteers willing to donate their computers' resources to the project. They can download the client from M4's Web site ( www.bytereef.org/m4_project.html).This Enigma portable encryption encoder/decoder is like the ones the German military used during World War II. A recent distributed-computing project decoded two intercepted German Navy Enigma messages that had remained unbroken since 1942.