No abstract

Stream processing platforms allow applications to analyse incoming data continuously. Several use cases exist that make use of these capabilities, ranging from monitoring of physical infrastructures to pre selecting video surveillance feeds for human inspection. It is difficult to predict how much computing resources are needed for these stream processing platforms, because the volume and velocity of input data may vary over time. The open source Apache Storm software provides a framework for developers to build processing applications that use the computing resources of all machines within an established cluster. Because of the varying processing needs of such applications, the platform should be able to automatically grow and shrink as needed. Unfortunately, the current Storm platform does not provide this capability. In this paper we describe the design and implementation of a tool that monitors several aspects of the Storm platform, the applications running on top of it, and external systems such as queues and databases. Based on this information, the tool decides whether extra servers are needed or machines may be decommissioned from the cluster, and then acts upon it by actually creating new virtual machines or shutting them down.

Abstract-Cloud computing enables on-demand access to a shared pool of IT resources. In the case of Infrastructure as a Service (IaaS), the cloud user typically acquires Virtual Machines (VMs) from the provider. It is up to the user to decide at what time and for how long they want to use these VMs. Because of the pay-per-use nature of most clouds, there is a strong incentive to use as few resources as possible and release them quickly when they are no longer needed. Every step of the deployment process, i.e., acquiring VMs, creating network links, and installing, configuring and starting software components on them, should therefore be as fast as possible. The amount of time the deployment process takes can be influenced by the user by performing some steps in parallel or using timing knowledge of previous deployments. This paper presents four different strategies for deploying applications on cloud computing infrastructures. Performance measurements of application deployments on three public IaaS clouds are used to show the speed differences between these strategies.

No abstract

Software agent-based strategies using micro-economic theory like PowerMatcher[1] have been utilized to coordinate demand and supply matching for electricity. Virtual power plants (VPPs) using these strategies have been tested in living lab environments on a scale of up to hundreds of households. So far, the coordination configuration of a VPP is fixed in these settings. The DREAM [2] framework architecture uses heterarchies to make parts of a VPP flexible in coordination strategy depending on the current operational grid status. In this way, a sub-VPP, serving one coordination objective, can decouple from and couple to an overarching VPP with another coordination objective dynamically. In this paper a grid congestion simulation with an overarching VPP coordinating demand and supply for electricity market optimization [3] and a sub-VPP reacting to a heat-pump congestion event in winter and a PV overproduction event in summer is described. The simulation was run in a static simulator [4]. The LVsegment consisted of 'flameless' residential areas with well-insulated homes with primarily heat pumps for heating and some renovated homes with local gas-fired co-generators of heat and electricity. Households additionally had solar cells, batteries and EV charging units. The goal of the additional coordination sub-VPP was to solve grid stability issues like congestion due to heat pump loads in winter and overproduction by PV in summer in this physical part locally, while the rest of the cluster remained unaffected and still optimizing for the commercial goal. The results were analyzed in terms of infringement of comfort parameters and performance in adapting the flexible load and generation. It appeared, substantial load shedding and load shifting of devices is possible to show the synergy in solving the grid stability issues evenly sharing the discomfort to the individual heating devices. By changing their charging strategy, the new algorithm also showed heat storage and electricity storage devices providing additional support.