Situation You have got an SAP BW system with a number of (software) servers. You want to distribute the workload of the info masses and different information warehouse administration actions in a means that matches your wants finest. This might imply that you just want to have all processes distributed throughout all accessible servers or that you just want to have one devoted server for these processes. 2 Introduction SAP makes use of the phrases occasion and software server synonymously. So as to keep away from misunderstandings we use the time period occasion for an SAP occasion (software server) on this doc.
For a bodily machine we use the time period server. A number of the settings described on this doc are finished on an occasion degree, some on a server degree. In the event you don’t have a number of cases (of the identical SAP system) on one server you don’t have to attract this distinction between occasion and server when studying this doc. There are a number of features and settings within the space of load balancing supplied by the idea system (Internet Utility Server). Nonetheless, these have been designed primarily for SAP’s ERP system.
Customizing these options for optimum use with SAP BW requires additional concerns. The challenges introduced with information load processing originate from the truth that many pretty lengthy working processes will be began virtually simultanesouly. The usual SAP load balancing strategy takes the standard of the cases into consideration when distributing the load. This high quality is evaluated in common intervals (5 minutes by default). Inside one interval a variety of parallel processes could also be began on the most effective occasion, utilizing a variety of work processes whereas the opposite cases are idle.
An optimum distribution of BW OLAP workload or information load useful resource consumption can not readily be achieved with this normal technique. With out enough planning, and underneath heavy workload (peak) circumstances, the dangers can enhance that turns into a bottleneck; a restricted variety of servers can turn into saturated with processes consuming assets, and efficiency (and stability) can probably undergo considerably. A profitable load balancing strategy optimally makes use of the assets which have been allotted to the BW system.
Notice that this dialogue assumes that an enough sizing has been carried out to correctly measurement the SAP BW system (see SAP Service Market alias “quicksizer” for extra data). This doc describes load balancing approaches for typical SAP BW actions. Generally these actions course of giant quantities of knowledge. Data (inside one course of) is cut up into packages and might thus be processed in parallel on one or throughout a number of servers or cases. Then again, a number of processes can run in parallel on one or on a number of servers or cases.
Which means we will have parallel processing (and consequently obtain load balancing) each inside one course of and throughout processes. In our examples we are going to use a system referred to as XXX as SAP BW system and a system referred to as YYY as an SAP supply system of XXX. Throughout information load processing, information is extracted from the supply system and despatched to the goal SAP BW system. Different load processes contain the SAP BW system as supply system, in addition to the goal system (for instance, DataMarts, activation of knowledge in ODS objects). The cases and servers on XXX are as follows, the server ab1234 being the database server:
