I spoke about the Vestas Project yesterday (2012.03.15) at the event called
"A New St. Louis? Powering Smart Cities with HPC, Big Data and Cloud". The event
was held at Missouri History Museum and organzied by STLhpc.net, the brain child
of Gary Stiehr of Washington University.
The project showcases how both Big Data and HPC converge to tackle both data-intensive and compute-intensive problems that are becoming more common in public and private data analytics. As a starter, the project will build a supercomputer capable of analyzing diverse and large weather data sets reaching 20-plus petabytes and reducing the time-to-results from weeks to less than an hour.
On October 24th 2011, IBM and Danish wind turbine manufacturer Vestas announced that a supercomputer code name "Firestorm" will analyze petabytes of data to optimize the placement and maximize energy output of Vestas turbines.
"Firestorm" will crunch through weather reports, moon and tidal phase, geospatial sensor data, satellite images, and deforestation maps to generate the best placement of turbines.
"Firestorm" will also help Vestas to see into the future and prescribe solutions. Software on the system will be used by anaylsts to model and research weather in predicting future performance. Vestas engineers will run other softwre to figure out the best time to do maintenance of the turbines.
Predicting energy output of turbines is vitally important to project developers who put up money for wind farms with an expectation of selling a certain amount of energy to the grid. Although wind power is growing in many places around the world, project developers are seeking better methods, including better wind speed measurement, to better match expected and actual performance.
The IBM software package is called BigInsight and it took four years to develop. BigInsight will run on the open-source Apache Hadoop framework for parallel processing of very large data sets. The software provides a framework for large scale parallel processing and scalable storage for terabyte to petabytes-level data plus the ability to enable "what-if" scenarios with its BigSheets component. BigInsights is part of IBM's Big Data software platform, which includes InfoSphere Streams software that analyzes data coming into an organization and in real time and monitors it for any changes that may signify a new pattern or trend.
A bit more on the "Firestorm" itself: the supercomputer has 1,222 connected, workload optimized System x iDataPlex servers and is capable of 150 trillion calculations per second -- equivalent to 30 million calculations per Danish citizen per second. Firestorm is #53 on the Top500 list of the world’s fastest supercomputers and the third largest commercial system on the list.
One of my colleagues from the IBM Worldwide Deep Computing team, Scott D, got to architect the system and worked with Vestas closely on the project. He's a lucky guy!
Links:
The project showcases how both Big Data and HPC converge to tackle both data-intensive and compute-intensive problems that are becoming more common in public and private data analytics. As a starter, the project will build a supercomputer capable of analyzing diverse and large weather data sets reaching 20-plus petabytes and reducing the time-to-results from weeks to less than an hour.
On October 24th 2011, IBM and Danish wind turbine manufacturer Vestas announced that a supercomputer code name "Firestorm" will analyze petabytes of data to optimize the placement and maximize energy output of Vestas turbines.
"Firestorm" will crunch through weather reports, moon and tidal phase, geospatial sensor data, satellite images, and deforestation maps to generate the best placement of turbines.
"Firestorm" will also help Vestas to see into the future and prescribe solutions. Software on the system will be used by anaylsts to model and research weather in predicting future performance. Vestas engineers will run other softwre to figure out the best time to do maintenance of the turbines.
Predicting energy output of turbines is vitally important to project developers who put up money for wind farms with an expectation of selling a certain amount of energy to the grid. Although wind power is growing in many places around the world, project developers are seeking better methods, including better wind speed measurement, to better match expected and actual performance.
The IBM software package is called BigInsight and it took four years to develop. BigInsight will run on the open-source Apache Hadoop framework for parallel processing of very large data sets. The software provides a framework for large scale parallel processing and scalable storage for terabyte to petabytes-level data plus the ability to enable "what-if" scenarios with its BigSheets component. BigInsights is part of IBM's Big Data software platform, which includes InfoSphere Streams software that analyzes data coming into an organization and in real time and monitors it for any changes that may signify a new pattern or trend.
A bit more on the "Firestorm" itself: the supercomputer has 1,222 connected, workload optimized System x iDataPlex servers and is capable of 150 trillion calculations per second -- equivalent to 30 million calculations per Danish citizen per second. Firestorm is #53 on the Top500 list of the world’s fastest supercomputers and the third largest commercial system on the list.
One of my colleagues from the IBM Worldwide Deep Computing team, Scott D, got to architect the system and worked with Vestas closely on the project. He's a lucky guy!
Links:
No comments:
Post a Comment