Optimizing Data Transfer in Telecommunications
Whether in medical technology, aerospace research or computation of graphics for film sequences, information exchange is an essential part of the way we now live and work. Complex systems such as heterogeneous network architectures can be individually mapped and realistically simulated using software and hardware developed by the Fraunhofer Heinrich Hertz Institute.
The Open Source Software QUADRIGA – short for Quasi Deterministic Radio Channel Generator – answers questions such as where the best reception can be found or where the next transmitter antenna should be located. By computing mobile wireless channels both inside and outside buildings, the software facilitates efficient planning of future radio networks.
Advantages of QUADRIGA
- No memory bottlenecks
- Accelerated data transmission
- Simultaneous access to all data
The statistical description of a channel serves as the basis for the modelling of a real world environment. Algorithms for reduction of interference in the transmission channel and for tolerance of poor reception can be evaluated and tested in an artifical environment. However, this requires extensive information theory calculations which must all take account of a huge number of different factors. Quadriga is an application which natively runs on the Fraunhofer Heinrich Hertz Institute’s High Performance Cluster Computing Platform.
High Performance Cluster Computing Platform
High Performance Computing (HPC) is an aggregation of computing power which generates performance levels many times higher than a single desktop. HPC refers to the use of high performance computers for the computation of complex procedures in such fields as:
- Medical technology
- Aerospace research
- Encoding of film sequences
- Meteorological models
- Virtual Reality
- Simulation technology
HPC offers a level of computing power far superior to that of conventional processors. The High Performance Cluster Computing Platform (HPCC) platform at the Fraunhofer Heinrich Hertz Institute enables seamless collaboration between hundreds or thousands of processor cores. The system’s memory has been adapted to eliminate bottlenecks in access to memory and to speed up the whole system while the simultaneous access to all data is enabled through a shared distributed file system with high speed interconnection. The workflow across the system makes effective use of system resources by distributing them efficiently among current processes.