Scalable Ray Launching Approach

The RPS ray launching approach allows for unlimited parallelization of the calculations with minimal overhead over as many workstations as are available in a TCP/IP LAN. Moreover, advanced load balancing between the workstations ensures that the computational power is always exploited to a maximum. The Figure below gives an estimate on how drastically the runtime can be reduced by distributing the computations over 10 equal workstations (middle) compared with a local simulation (left). In a reference installation on a fully equipped 56 XEON™ processor blade server, the system showed unprecedented performance as indicated to the right.

The parallelized ray launching approach of RPS is realized in a client-server architecture. The RPS Client application controls the simulation and interacts with the RPS Servers that are distributed over the LAN.

Scalable RNP Solution

In radio network planning (RNP), there are various tasks that require a frequent recalculation of the radio channel e.g. for new sites or changed base station configurations (antenna height etc.). Thus for a network operator, it is more suitable to deploy several RPS Client applications (graphical user interface incl. simulation control) in parallel at several seats. Each of the clients can control several RPS Servers.

RPS supports this architecture by providing a special simulation control that can share RPS Servers between an arbitrary number of RPS Clients. The servers are running in the background on different workstations, thus forming a pool of available RPS compute power. All currently idle servers can be accessed from each of the clients at any time, yielding an optimal exploitation of the shared compute pool.

RPS Server applications are available today for Windows 95/98 and NT 4.0 (SP3 or higher), NT 5 (Windows 2000), or Windows XP (SP2 or higher) operating systems, as well as for x86-Linux 2.x systems.