Realizing Elastic Processes with ViePEP

Online business processes are faced with varying workloads that require agile deployment of computing resources. Elastic processes leverage the on-demand provisioning ability of Cloud Computing to allocate and de-allocate resources as required to deal with shifting demand. To realize elastic processes, it is necessary to track the current and future system landscape, monitor the process execution, reason about how to utilize resources in an optimal way, and carry out the necessary actions (e.g., start/stop servers, move services). Traditional Business Process Management Systems (BPMS) do not consider such needs of elastic process. Within this demo, we present ViePEP, a research BPMS able to execute and monitor resource-, costand QoS-elastic, service-based workflows and optimize the overall system landscape based on a reasoning of the non-functional requirements of current and forthcoming elastic processes. 1 Significance to the Field Resource-intensive tasks are nowadays not only common within scientific workflows, but are also getting more and more common in business processes.. For example, computeand data-intensive analytical processes are found in the finance industry and in managing smart grids in the energy industry. In the latter case, data from a very large number of sensors needs to be gathered, processed and stored in real-time in order to offer consumers consumption reports or even guarantee grid stability [5]. As the number of active sensors differs during a day, the amount of data also fluctuates to a very large extent. Furthermore, certain processes or process steps are permitted to be postponed to the future, while others need to be carried out immediately. In such a scenario, the permanent provisioning of IT capacity able to handle peak system loads is obviously not the best solution, as the capacities will not be utilized most of the time. With the advent of Cloud Computing, organizations nowadays have got a much more cost-savvy alternative which allows them to make use of computing resources in an on-demand, utility-like fashion [1]. 1 In the following, we will also make use of the term “workflows” to name the automated parts of business processes. A. Ghose et al. (Eds.): ICSOC 2012, LNCS 7759, pp. 439–442, 2013. c © Springer-Verlag Berlin Heidelberg 2013 440 S. Schulte et al. To the best of our knowledge, so far, little effort has been put into the investigation of methods and tools to integrate automated process execution and Cloud Computing in order to realize so-called elastic processes [2]. Nevertheless, scalability and cost-effective allocation of single tasks and applications have been observed by many researchers, e.g., [4]. So far, applications are mostly regarded in isolation, i.e., a process perspective across utilized resources (software, hardware, humans) is typically not applied. While Cloud resources have been used for executing scientific workflows [3], Service Level Agreements (SLAs) are typically not as much a concern in this domain as they are for business processes. In our experience, there is a lack of a BPMS able to carry out many service-based workflows in parallel, estimate their current and future resource demand under given Quality of Service (QoS) constraints, and allocate resources dynamically to meet their individual SLAs. This needs analysing the process to discover which of its steps determine the performance of its execution and prioritising them, allocating resources to services to address demand and balancing load on the resources by moving services between them. Within this demo paper, we will present our Vienna Platform for Elastic Processes (ViePEP) [6], which is a research-driven, prototypical BPMS capable to execute elastic processes, monitor the current utilization of invoked resources as well as reason about future resource demands, and carry out necessary actions.