Faris Nizamic

Towards a Robustness Evaluation Framework for BPEL Engines

The selection of the best fitting process engine for a specific project requires the evaluation of engines according to various requirements. We focus on the non-functional requirement robustness, which is critical in production environments but hard to determine. Thus, we propose an evaluation framework to reveal important robustness criteria of process engines. In this work, we focus on message robustness, i.e., The ability to handle the receipt of invalid messages appropriately. In a case study comprising five open source BPEL engines, we determine message robustness by injecting faults into robustly designed processes as a reply to a previously sent request from an external virtual service and assert their behavior. The results show that the degree of message robustness significantly differs, hence, robustly designed processes do not necessarily lead to robust runtime behavior, the selected engines still play a major role.

Power management of personal computers based on user behaviour

It has been shown that up to 64 percent of personal computers in office buildings are left running during after-hours. Enabling power management options such as sleep mode is a straightforward method to reduce the energy consumption of computers. However, choosing the right timeout can be challenging. A sleep timeout which is too low leads to discomfort, whereas a timeout which is too high results in poor energy saving efficiency. Having the users choose their own sleep timeout is not viable as research shows that most users disable the sleep timeout completely, or choose a suboptimal timeout. Unlike existing context based power management systems which use predefined rules, we propose a solution which can determine a personalized sleep timeout for any point in time solely based on the users behaviour. We propose multiple models which have the goal of maximizing the energy savings while minimizing discomfort. The models are tested on the computers of employees of the University of Groningen over several weeks. We analyse the results of the experiments and determine which model performs best. We can potentially save between 4.02 and 17.17 kWh per computer per year, depending on the model that used.

GreenMind-An Architecture and Realization for Energy Smart Buildings

Sustainability and energy-efficiency are receiving increasing attention. Existing buildings are responsible for more than 40% of the worlds total primary energy consumption. Current building management systems fail to reduce unnecessary energy consumption and preserve to user comfort at the same time mainly because they are unable to cope with dynamical changes caused by user’s interaction with the environment. To cope with this dynamicity, we propose a software architecture for energy smart buildings that includes a set of concrete software solutions that tackle energy consumption sub-systems; i.e., heating, ventilation and air conditioning (HVAC), lighting, workstations and other appliances subsystems, in order to save significant amount of energy whilst preserving user comfort. Experimental results carried out in a 12.000 square meter building of the University of Groningen show that our proposed solutions are able to save up to 56% of electricity used for lighting, at least 20% of electricity used for heating while the savings from controlling workstation as well as other appliances are 33% and 10%, respectively.

Planning meets activity recognition: Service coordination for intelligent buildings

Abstract Building managers need effective tools to improve occupants’ experiences considering constraints of energy efficiency. Current building management systems are limited to coordinating device services in simple and prefixed situations. Think of an office with lights offering services, such as turn on a light, which are invoked by the system to automatically control the lights. In spite of the evident potential for energy saving, the office occupants often end up in the dark, they have too much light when working with computers, or unnecessary lights are turned on. The office is thus not aware of the occupants’ presence nor anticipates their activities. Our proposal is to coordinate services while anticipating occupant activities with sufficient accuracy. Finding and composing services that will support occupant activities is however a complex problem. The high number of services, the continuous transformation of buildings, and the various building standards imply a search through a vast number of possible contextual situations every time occupants perform activities. Our solution to this building coordination problem is based on Hierarchical Task Network (HTN) planning in combination with activity recognition. While HTN planning provides powerful means for composing services automatically, activity recognition is needed to identify occupant activities as soon as they occur. The output of this combination is a sequence of services that needs to be executed under the uncertainty of building environments. Our solution supports continuous context changes and service failures by using an advanced orchestration strategy. We design, implement and deploy a system in two cases, namely offices and a restaurant, in our own office building at the University of Groningen. We show energy savings in the order of 80% when compared to manual control in both cases, and 60% when compared to using only movement sensors. Moreover, we show that one can save a figure of €600 annually for the electricity costs of the restaurant. We use a survey to evaluate the experience of restaurant occupants. The majority of them are satisfied with the solution and find it useful. Finally, the technical evaluation provides insights into the efficiency of our system.

Testing of Distributed Service-Oriented Systems

We are experiencing an exponential growth of devices connected to the Internet and services offered through the web. Today, we are just a few mobile-clicks away using services which enormously simplify our life. Just think of how we are paying our bills, recharging our mobile pre-paid account, or how we buy tickets for the events we want to attend. It is all being done through web services. This increasing reliance on distributed service-oriented systems provided through the web places a high expectation on their reliability. To keep up with this growing trend that is embracing changes on a daily basis, the software development of the services has to be rapid and at the same time leaving not much space for software errors or failures.

Cloud Ready Applications Composed via HTN Planning

Modern software applications are increasingly deployed and distributed on infrastructures in the Cloud, and then offered as a service. Before the deployment process happens, these applications are being manually - or with some predefined scripts - composed from various smaller interdependent components. With the increase in demand for, and complexity of applications, the composition process becomes an arduous task often associated with errors and a suboptimal use of computer resources. To alleviate such a process, we introduce an approach that uses planning to automatically and dynamically compose applications ready for Cloud deployment. The industry may benefit from using automated planning in terms of support for product variability, sophisticated search in large spaces, fault tolerance, near-optimal deployment plans, etc. Our approach is based on Hierarchical Task Network (HTN) planning as it supports rich domain knowledge, component modularity, hierarchical representation of causality, and speed of computation. We describe a deployment using a formal component model for the Cloud, and we propose a way to define and solve an HTN planning problem from the deployment one. We employ an existing HTN planner to experimentally evaluate the feasibility of our approach.