Load Balancing for Loading and Data Warehouse Processes
If you are using a BW system with multiple application servers, you can balance the workload of your loading processes and other Data Warehouse processes in the way that best meets your requirements. You might want to balance all processes over all available application servers, or use a certain application server for the processes.
In the following sections, the application servers are referred to as instances, and the physical units are referred to as hosts.
Without adequate planning, there is an increased risk that a high system load will cause bottle necks in the hardware. With processes that make intensive use of resources, some servers can reach their limits. As a result, the performance (and stability) of the system can be seriously impaired. A successful strategy for load balancing makes optimum use of the hardware available for the BW system. The following sections describe possible methods for ensuring load balancing of the loading processes and typical Data Warehouse processes in the BW system: These processes generally process large quantities of data. The data (in a process) is split into packages and can thus be processed in parallel on one or more hosts and instances. On the other hand, various processes can be processed in parallel on one or more hosts and instances. This means that parallel processing and load balancing are possible both in a single process and across multiple processes.
In the following information, note that we assume that adequate sizing has already been performed for the BW system.
You have the following options for load balancing while processing loading processes and Data Warehouse processes: