Lisa Maile

This research project deals with the application of quality of service guarantees in Time-Sensitive Networking, in particular using Network Calculus. Real-time systems are increasingly required in industry, e.g. the automotive, automation or entertainment industries. Classical Ethernet, however, does not guarantee real-time performance, which leads the Time-Sensitive Networking Task Group (IEEE 802.1) to develop standards for real-time data transmission over Ethernet networks. These standards are summarized under the term Time-Sensitive Networking (TSN). Within the scope of this research project, the application of Network Calculus for TSN is now being investigated. Network Calculus (NC) is a system theory for deterministic performance evaluation. It uses mathematical methods to provide performance guarantees for communication systems. NC can help evaluate TSN's real-time properties, meet required latency limits, and provide insight into the optimal configuration of networks. It also enables buffer sizing and can evaluate existing or new scheduling algorithms.

Network calculus (NC) is a system theory for deterministic performanceevaluation. It uses mathematical methods to provide performanceguarantees for communication systems. It can be applied in thedesign phase of future systems as well as the analysis of existingsystems. In real-time systems, the timeliness of events plays animportant role. Therefore, the classical performance evaluation based onstochastic methods that result in (stochastic) expectation values, i.e.mean values, has to be extended by mathematical tools producingguaranteed bounds for worst case scenarios. Network calculus allows toobtain upper bounds for end-to-end delays for one nodes or aseries of nodes within a network, upper bounds for the required bufferspace and bounds for the output flow.These analytic performance bounds characterize the worst-case behaviorof traffic flows and allow dimensioning the corresponding systems.

Currently, we study the applicability of NC for multiplexed flows, inparticular when the FIFO property cannot be assumed at the merging ofindividual flows. The aggregation of data flows plays an important rolein modelling the multiplexing scheme. We apply NC for performanceevaluation both of aggregate multiplexing at one node and atconcatenation of aggregated multiple nodes in different scenarios.
We have successfully introduced network calculus methods in thefield of internal automotive communication systems in industrialapplications. Embedded in-car networks need to fulfill hardreal-time constraints. While TDMA-based access schemes in FlexRayguarantee that certain bound can be met, statistical multiplexingin CAN networks only allows to calculate bounds for the highestpriority messages. By applying network calculus, we obtained boundsfor all priority classes without the need to specify a concretescheduling of the messages. Upper bounds for the amount of datathat arrives at each network node are enough to determine hardbounds for the end-to-end delay in CAN networks.

Another field of application is industrial communication.Factory automation often also requires hard real-time boundsfor the end-to-end delay of messages. The use of Ethernet withpriority tagging allows cost-efficient implementation offactory automation systems. But without stringent planningof the network, the required bounds on the end-to-end delaycannot be guaranteed. Network calculus allows to obtain therequired bounds when applied in the planning phase of thenetwork. It also allows to dimension the buffers of nodes,e.g. of industrial Ethernet switches. Nowadays, some ofthe users of industrial Ethernet need to integratenon-real-time products like web cams and remote operationterminals into existing networks. Withoutadditional analysis, the additional traffic caused by devicesthat do not require hard real-time constraints willcause a violation of the bounds for the delay and bufferspace for real-time traffic. By taking into account thisnon-real-time traffic in network calculus and by applyingtraffic shaping for the non-real-time flows allows todimension the network so that all bounds are met.Network calculus is currently integrated into an existingautomated industrial network planning tool.