Modeling of Computer Networks Using SAP HANA Smart Data Streaming

Abstract

This article presents an algorithm for transforming a mathematical model into a form accepted by the SAP HANA Smart Data Streaming (SDS). The implementation is intended to enable quick and easy processing of numerical data, generated from the model to evaluate the state of selected scenarios and network topologies. The mathematical model used is a fluid-flow approximation algorithm, describing changes in the dynamics of data transmission and changes in the queue length of packets waiting in network nodes over time. The adopted methodology assumes the use of an input stream, which transports the tuples with timestamps, as a trigger for determining the next steps of the model. Parameter values are stored in the column tables in the SAP HANA Platform database. In addition, the article proposes the extension of the model by the possibility of automatic and manual switching of parameters in the network. The purpose of the extension is to dynamically increase or reduce hardware resources for each of the network nodes, depending on the load of the given node. The modification introduces four additional thresholds, which are responsible for controlling the switching between operating modes: normal, high performance and energy saving. The introduction of modifications provides model’s parameters reloads during numerical calculations. It also allows optimized resources utilization for network operations, with a small reduction in average flow throughput and a slight increase in the router’s buffer occupancy.