The amount of data and number of sources that can be drawn to inform medical design is expanding exponentially. In addition to the explosion of web-based information, sensor-based wearable medical devices offer a treasure trove of health and wellness-related data. All of this sets the stage for unprecedented leaps forward in healthcare. There’s only one problem: traditional data infrastructure can’t process at that volume and speed. As we anticipate most life-saving advances will involve processing enormous amounts of data from numerous streams in real time, this creates a bit of an impedance to progress. Between this and increasingly stringent record-keeping regulations, healthcare data demands are piling up, and traditional infrastructure can’t keep pace.
In-Memory Computing (IMC) technology, however, offers a way to achieve performance orders of magnitudes faster than standard methods. IMC involves distributed processing of data in the computer’s memory, and was once primarily in the domain of academia. Recently, however, a few companies have begun offering commercially viable platforms, and now IMC can be performed on commodity hardware and accessed through open source software at a cost comparable to traditional processing.
The implications of this kind of computing performance for medical design are inestimable. Consider, for example, Medical Image Processing. IMC technology can enable real-time processing and analysis of images from PET scans to show organ and tissue health; or in detecting and diagnosing diseases through MRI or CT scans. Rather than waiting days or weeks for results, physicians will identify problems and prescribe treatments immediately.
Additionally, consider the advances in wearable medical devices. While we’re seeing some amazing breakthroughs, these devices are currently limited by processing power, and our ability to turn the data they generate into actionable information is hamstrung. However, by using computing technology 100x faster than what’s even theoretically possible with flash-based storage, the limits are removed. With IMC, it is very conceivable that within just a few years we could see apps that monitor health signs on a person 24/7, interact with data from their medical records, diagnose problems and alert their physician immediately when they need help.
If this sounds far-fetched, research is currently being done. For example, Portland State University’s program, "Computing with Biomolecules: From Network Motifs to Complex and Adaptive Systems," leverages IMC from GridGain Systems to achieve one of the most advanced applications of computational chemistry to date. The project’s goal is to build adaptive learning systems for pathogen detection, medical diagnosis and therapeutic interventions. Eventually, developments will bring biocompatible bio-molecular computers capable of embedding a living cell or body. In addition to enabling massive simulations that are beyond the capabilities of traditional infrastructure, IMC technology facilitates collaboration between teams on the National Science Foundation (NSF) project at both the University of New Mexico and Columbia University, allowing team members to log in and run simulations on each other’s clusters.
The pieces are in place for incredible advances in healthcare. IMC provides the final component necessary to harness the growing amount of data we are able to gather, and turn it into insights, diagnoses, and treatments.
For more information, visit GridGain.