Intel Sees More ‘Aware’ Technologies Which Can Adapt To People In Next Decade
"A user-aware platform will be any device that can take care of itself, knows who we are, where we are, and tries to anticipate what we want done," Rattner said. "They will need digital senses to be aware of their surroundings and what they are doing. They will also need new levels of intelligence to understand our needs and collaborate with other electronics to take action on our behalf while doing no harm in the process."
User-Aware Systems Will Require New Levels of Intelligence
User-aware platforms will place new demands on the way the industry develops hardware, software, services and interfaces. Intel is already evolving its processor architectures toward supercomputer-like performance by employing multiple processing cores in each processor (which is like adding multiple "brains" to a computer). To enable platforms that are more aware and help meet people’s needs in future, Intel envisions putting tens to even hundreds of energy-efficient computing cores inside a single processor.
Each chip will be capable of dynamically assigning individual or clusters of processing cores along with the necessary memory and bandwidth to specific tasks such as seeing, listening, network security, gaming, and understanding commands. These future platforms will use virtualization software to put protective walls around each task’s dedicated slice of computing resources so they run better and don’t interfere with other applications.
User-aware platforms will use this new level of intelligence to manage multiple tasks and sources of input (video to see with, audio to hear and speak, sensors to feel, storage to remember, the network and radios to connect with the Internet and other devices) and then apply machine learning to simplify our lives. To illustrate this point, Rattner demonstrated a research project called Diamond, a more intuitive image search application under development by researchers at Intel and Carnegie Mellon University in Pittsburgh.
"Imagine if all the pages of all the medical books in the library of congress fell out on the floor and you had to search through them for a specific image of a cancer cell. No file names, no folders, just piles of images; this is called searching non-indexed data," Rattner said. "Running simultaneously across several computers, Diamond uses advances in computer vision and machine learning to search through data the way people do, first by studying what the desired image ‘looks like’ – its shape, color, contents – and then finding the closet matches. It is an initial attempt to do for complex data what spreadsheets have done for numbers."
Continuing Platform Innovations Required To Make Systems User-Aware
For the user-aware platform to take care of itself it must be able to sense its surroundings and then respond to threats and problems that endanger its ability to function. To achieve this Intel is investigating the use of sensor technologies combined with dedicated logic within systems. One example is installing heat sensors inside servers to monitor for overheating. They can trigger the intelligent reallocation of workloads among hundreds of systems inside a data center to avoid data lose or system failure.
Rattner demonstrated how network awareness built into individual devices could be used to stop worm attacks at the first infected system. Aggressive worm attacks can propagate throughout a network jumping from system to system faster than people can physically respond. Circuit Breaker is an Intel research project that stops novel worms in their tracks by enabling each device to intelligently monitor the health of its own network traffic. Circuit Breaker looks for abnormalities in a range of system network behaviors and isolates its system from the network at the first sign of an attack. In this way, "self-policing" systems can form a last-line of defense against the spread of a worm.
Perhaps the most difficult challenge of building systems that intuitively respond to our ever changing needs is helping them understand context: the "who," "what" and "where" of our lives. Location-aware computing technology could help systems more intuitively respond to requests, such as downloading different types of music to the family car, kitchen or cell phone based on a person’s tastes. Or, it could alert a factory worker to specific maintenance needs of different equipment based on proximity, safety conditions and the worker’s level of training. And, it should do this in an appropriate, privacy observant manner.
Intel will work closely with the industry – original equipment manufacturers (OEMs), independent software vendors (ISVs) and developers – on a number of fronts to make this vision of user-aware computing a reality and to bring better, more intelligent products to people around the world that are useful to them where they live.