Tommy Baumann

Analysis and Evaluation of the German Toll System Using a Holistic Executable Specification

The European Electronic Toll Service (EETS) allows different possibilities for the overall system design of interoperable toll systems. We study the effect of the overall system design on the operational costs using a model based, holistic, executable specification. The model is set up to give an accurate description of the interdependences as seen in the German toll system. Two different scenarios are simulated and compared. The first scenario simulates the on-board toll calculation (as used in the German toll system), the second scenario corresponds to backend toll processing. The simulation gives an accurate figure for the operational costs associated with this design decision. A comparison of the cost of financing (of not yet processed toll fees) with the cost of communications is derived from the simulation results and allows choosing the right system architecture and system parameterization according to minimal operational costs.

Network Resource Usage of the German Toll System: Lessons from a Realistic Simulation Model

We are using a detailed realistic discrete event simulation model of the automatic German toll system to study resource usage of the mobile data network for various system configurations. Starting with an existing simulation model [1] we describe the driving patterns used in the simulation and the functionality added to the simulation model. Next we explain enhancements in simulation performance implemented to reach simulation runs covering 16 weeks of simulation time for a vehicle fleet of 700.000 heavy trucks. Using the simulation model we analyze the number of simultaneous mobile data network connections which is strongly affected by a single system parameter. The realistic configuration is used to study the speed of rolling out map data updates across the vehicle fleet.

Using Simulations to Study the Efficiency of Update Control Protocols

Taking an existing large-scale realistic discrete event simulation model [1, 2] of a mobile distributed system with more than 700 000 units we investigate the efficiency of control protocols for the software and data updates of on-board-units (OBUs). Adding four different closed-loop control protocols to the existing simulation model we explain and investigate the technical implications of the control protocols: OBUs that allow incoming network connections can be reached at any time (Always-On or Sometimes-On). If the OBU is designed without a TCP/IP server, signaling is implemented either as polling (Sometimes-On with polling) or using a dedicated GSM signaling channel (On-Demand-On). The efficiency is discussed as a combination of metrics gathered during simulation runs and compared with the OBU controlled update logic. These simulation results are subsequently used as input for the system development process.

Using parameter optimization to calibrate a model of user interaction

Simulation models of real-world distributed systems depend both on the accuracy of the underlying model and the interaction between user and system. The user interaction is typically modeled as stochastic process depending on parameters and distributions describing the actual usage. Accurate data is often not available and (manual) assumptions are necessary. Taking an existing large-scale simulation model of the German tolling system we discuss the use of a genetic optimization algorithm for calibrating the simulation model.

Simulation Driven Development of the German Toll System – Simulation Performance at the Kernel and Application Level

Simulation driven development - the idea of using simulation models as executable system specification in any phase of the system development process [4] - depends on the performance of the simulation model and execution framework. We study the performance issues of an existing large-scale simulation model of the German toll system using a discrete-event simulation (DES) model. The article first introduces the German toll system and the simulation framework developed to analyze the systems’ behavior. To address the simulation performance the article describes a number of common performance limitations of several commercial and non-commercialDES simulation kernels. These performance limitations are addressed in kernel-level benchmarks. At the application-level a DES implementation of the German toll system is used to compare two commercial DES tools and several optimizations are introduced both on the simulation model and kernel level to achieve the necessary performance for a detailed and realistic simulation of a fleet of 750 000 trucks.

Modeling Regional Reliability of 2G, 3G, and 4G Mobile Data Networks and Its Effect on the German Automatic Tolling System

Using an existing large-scale realistic discrete event simulation model of a mobile distributed system we model the infrastructure and dependencies of 2G, 3G and 4G mobile data networks. We investigate the effects of the different network generations and of transitioning connections between networks on the typical behaviour of the distributed system. In particular we enable the model to simulate regional outages of networks to predict the effects of shared back-end network components and the effects of mobile endpoints entering and leaving regional coverage.

Enhanced simulation performance through parallelization using a synthetic and a real-world simulation model

Taking an existing large-scale simulation model of the German toll system we identify possibilities for parallelization in order to enhance simulation performance. We transform parts of the model from its current serial implementation to a parallel implementation. Afterwards we evaluate the achieved performance enhancement and compare the results to a synthetic benchmark model.

Network-wide Measurement of TCP RTT in 2G Networks

We analyze existing server-side log data of a large scale automatic toll system to measure the TCP roundtrip-time (RTT) as experienced by the communication between the central system and the on-board units (OBUs) deployed for tolling heavy-goods vehicles. The RTT is estimated from passive monitoring by parsing server-side log files and aggregating fleet-wide statistics over time. Using this data we compare the characteristics of the four different types of OBU and the three GPRS (2G) networks used. We find the RTT data to be consistent with existing, smaller samples and extend the observed RTT range by an order of magnitude. The OBU types exhibit a markedly different behavior, most notably for long RTTs, and we find one of the 2G networks to ‘hum’ at 50 Hz and harmonics.

Modeling the GPRS Network Latency with a Double Pareto-lognormal or a Generalized Beta Distribution

Taking a newly collected large data set on the TCP connection termination latency in GPRS networks we try to identify the underlying statistical distribution. The data extends the observed latencies to large time scales necessitating a heavy-tail distribution. Many distributions work well for the main body of the data. However, the heavy tail of the distribution benefits from mixing different statistical distributions. We compare several distributions and find that the double Pareto-lognormal distribution and the generalized Beta distribution of the second kind fit the data equally well.

On the Necessity for High-availability Data Center Backends in a Distributed Wireless System

When business processes depend on the processing capabilities within a data center, the typical system architecture use a high-availability setup to maintain a high level of service. Faced with a specific machine-tomachine system consisting of many endpoints that collect and forward data to the data center we argue that the dependability of the overall system does not necessitate a high level of service for the data center components. Taking an existing discrete event simulation model of a distributed technical system we investigate and discuss the effects of prolonged outages of the data center on the major business processes of the system.