Message brokers often facilitate communication between data producers and consumers by adding variable-sized messages to ordered distributed queues. Our goal is to determine the number of consumers and consumer partition assignments needed to ensure that the data consumption rate matches the data production rate. We model this problem as a variable item size bin packing problem. As the production rate varies, new consumer–partition assignments are computed, potentially requiring the reallocation of partitions from one consumer to another. During reallocation, data in the queue are not read, leading to increased latency costs. To address this problem, we focus on the multiobjective optimization cost of minimizing the number of consumers and reducing latency. We introduce several heuristic algorithms and compare them to state-of-the-art heuristics. In our experimental setup, the proposed modified worst fit (MWF) heuristic achieves a 48% reduction, with a similar number of consumers, in comparison with the best fit decrease (BFD). In addition, MWF achieves a 99 t h percentile latency of 2.24 seconds compared with that of 364.66 with the approach by Kafka using the same number of consumers. Alternatively, to obtain a lower 99 t h percentile latency than our approach does, Kafka requires at least 60% more consumers than our method requires. • Modelling the consumer group autoscaling problem to consider variable message sizes and production rates. • Defining the Rscore metric to quantify latency between two consecutive consumer group assignments. • Exploring four variations of the bin packing approximation algorithms. • Implementing a fully functional framework for the consumer group autoscaling problem. • Evaluating the autoscaler response time when autoscaling the consumer group in a Kafka production infrastructure.

Developing and optimizing software applications for high performance and energy efficiency is a very challenging task, even when considering a single target machine. For instance, optimizing for multicore-based computing systems requires in-depth knowledge about programming languages, application programming interfaces (APIs), compilers, performance tuning tools, and computer architecture and organization. Many of the tasks of performance engineering methodologies require manual efforts and the use of different tools not always part of an integrated toolchain. This paper presents Pegasus, a performance engineering approach supported by a framework that consists of a source-to-source compiler, controlled and guided by strategies programmed in a Domain-Specific Language, and an autotuner. Pegasus is a holistic and versatile approach spanning various decision layers composing the software stack, and exploiting the system capabilities and workloads effectively through the use of runtime autotuning. The Pegasus approach helps developers by automating tasks regarding the efficient implementation of software applications in multicore computing systems. These tasks focus on application analysis, profiling, code transformations, and the integration of runtime autotuning. Pegasus allows developers to program their strategies or to automatically apply existing strategies to software applications in order to ensure the compliance of non-functional requirements, such as performance and energy efficiency. We show how to apply Pegasus and demonstrate its applicability and effectiveness in a complex case study, which includes tasks from a smart navigation system.

Virtual machine image retrieval a b s t r a c tInfrastructure-as-a-service (IaaS) Clouds concurrently accommodate diverse sets of user requests, requiring an efficient strategy for storing and retrieving virtual machine images (VMIs) at a large scale.The VMI storage management requires dealing with multiple VMIs, typically in the magnitude of gigabytes, which entails VMI sprawl issues hindering the elastic resource management and provisioning.Unfortunately, existing techniques to facilitate VMI management overlook VMI semantics (i.e at the level of base image and software packages), with either restricted possibility to identify and extract reusable functionalities or with higher VMI publishing and retrieval overheads.In this paper, we propose Expelliarmus, a novel VMI management system that helps to minimize VMI storage, publishing and retrieval overheads.To achieve this goal, Expelliarmus incorporates three complementary features.First, it models VMIs as semantic graphs to facilitate their similarity computation.Second, it provides a semantically-aware VMI decomposition and base image selection to extract and store non-redundant base image and software packages.Third, it assembles VMIs based on the required software packages upon user request.We evaluate Expelliarmus through a representative set of synthetic Cloud VMIs on a real test-bed.Experimental results show that our semantic-centric approach is able to optimize the repository size by 2.3 -22 times compared to state-of-the-art systems (e.g.IBM's Mirage and Hemera) with significant VMI publishing and slight retrieval performance improvement.

Infrastructure-as-a-service (IaaS) Clouds concurrently accommodate diverse sets of user requests, requiring an efficient strategy for storing and retrieving virtual machine images (VMIs) at a large scale. The VMI storage management require dealing with multiple VMIs, typically in the magnitude of gigabytes, which entails VMI sprawl issues hindering the elastic resource management and provisioning. Nevertheless, existing techniques to facilitate VMI management overlook VMI semantics (i.e at the level of base image and software packages) with either restricted possibility to identify and extract reusable functionalities or with higher VMI publish and retrieval overheads. In this paper, we design, implement and evaluate Expelliarmus, a novel VMI management system that helps to minimize storage, publish and retrieval overheads. To achieve this goal, Expelliarmus incorporates three complementary features. First, it makes use of VMIs modelled as semantic graphs to expedite the similarity computation between multiple VMIs. Second, Expelliarmus provides a semantic aware VMI decomposition and base image selection to extract and store non-redundant base image and software packages. Third, Expelliarmus can also assemble VMIs based on the required software packages upon user request. We evaluate Expelliarmus through a representative set of synthetic Cloud VMIs on the real test-bed. Experimental results show that our semantic-centric approach is able to optimize repository size by 2.2 - 16 times compared to state-of-the-art systems (e.g. IBM's Mirage and Hemera) with significant VMI publish and slight retrieval performance improvement.

editorial

Recently cloud computing has gained popularity among e-Science environments as a high performance computing platform. From the viewpoint of the system, applications can be submitted by users at any moment in time and with distinct QoS requirements. To achieve higher rates of successful applications attending to their QoS demands, an effective resource allocation (scheduling) strategy between workflow's tasks and available resources is required. Several algorithms have been proposed for QoS workflow scheduling, but most of them use search-based strategies that generally have a higher time complexity, making them less useful in realistic scenarios. In this paper, we present a heuristic scheduling algorithm with quadratic time complexity that considers two important constraints for QoS-based workflow scheduling, time and cost, named Deadline-Budget Workflow Scheduling (DBWS) for cloud environments. Performance evaluation of some well-known scientific workflows shows that the DBWS algorithm accomplishes both constraints with higher success rate in comparison to the current state-of-the-art heuristic-based approaches.

Modelling & Simulation (M&S) is broadly used in real scenarios where making physical modifications could be highly expensive. With the so-called Simulation Software-as-a-Service (SimSaaS), researchers could take advantage of the huge amount of resource that cloud computing provides. Even so, studying and analysing a problem through simulation may need several simulation tools, hence raising interoperability issues. Having this in mind, IEEE developed a standard for interoperability among simulators named High Level Architecture (HLA). Moreover, the multi-agent system approach has become recognised as a convenient approach for modelling and simulating complex systems. Despite all the recent works and acceptance of these technologies, there is still a great lack of work regarding synergies among them. This paper shows by means of a literature review this lack of work or, in other words, the sparse Cloud SimSaaS. The literature review and the resulting taxonomy are the main contributions of this paper, as they provide a research agenda illustrating future research opportunities and trends.

Consolidation consists in scheduling multiple virtual machines onto fewer servers in order to improve resource utilization and to reduce operational costs due to power consumption. However, virtualization technologies do not offer performance isolation, causing applications' slowdown. In this work, we propose a performance enforcing mechanism, composed of a slowdown estimator, and a interference- and power-aware scheduling algorithm. The slowdown estimator determines, based on noisy slowdown data samples obtained from state-of-the-art slowdown meters, if tasks will complete within their deadlines, rescheduling tasks if needed. When invoked, the scheduling algorithm builds performance and power aware virtual clusters to successfully execute the tasks. We conduct simulations injecting synthetic jobs which characteristics follow the last version of the Google Cloud tracelogs. The results indicate that our strategy can be efficiently integrated with state-of-the-art slowdown meters to fulfil contracted SLAs in real-world environments, while reducing operational costs in about 12%.

Efficient application scheduling algorithms are important for obtaining high performance in heterogeneous computing systems. In this paper, we present a novel list-based scheduling algorithm called Predict Earliest Finish Time (PEFT) for heterogeneous computing systems. The algorithm has the same time complexity as the state-of-the-art algorithm for the same purpose, that is, O(v2.p) for v tasks and p processors, but offers significant makespan improvements by introducing a look-ahead feature without increasing the time complexity associated with computation of an optimistic cost table (OCT). The calculated value is an optimistic cost because processor availability is not considered in the computation. Our algorithm is only based on an OCT that is used to rank tasks and for processor selection. The analysis and experiments based on randomly generated graphs with various characteristics and graphs of real-world applications show that the PEFT algorithm outperforms the state-of-the-art list-based algorithms for heterogeneous systems in terms of schedule length ratio, efficiency, and frequency of best results.

Cloud computing is increasingly being adopted in different scenarios, like social networking, business applications, scientific experiments, etc. Relying in virtualization technology, the construction of these computing environments targets improvements in the infrastructure, such as power-efficiency and fulfillment of users' SLA specifications. The methodology usually applied is packing all the virtual machines on the proper physical servers. However, failure occurrences in these networked computing systems can induce substantial negative impact on system performance, deviating the system from ours initial objectives. In this work, we propose adapted algorithms to dynamically map virtual machines to physical hosts, in order to improve cloud infrastructure power-efficiency, with low impact on users' required performance. Our decision making algorithms leverage proactive fault-tolerance techniques to deal with systems failures, allied with virtual machine technology to share nodes resources in an accurately and controlled manner. The results indicate that our algorithms perform better targeting power-efficiency and SLA fulfillment, in face of cloud infrastructure failures.

For most Heterogeneous Computing Systems (HCS) the completion time of an application is the most important requirement. Many applications are represented by a workflow that is therefore schedule in a HCS system. Recently, researchers have proposed algorithms for concurrent workflow scheduling in order to improve the execution time of several applications in a HCS system. Although, most of these algorithms were designed for static scheduling, that is all application must be submitted at the same time, there are a few algorithms, such as OWM (online workflow Management) and RANK_HYBD, that were presented for dealing with dynamic application scheduling. In this paper, we present a new algorithm for dynamic application scheduling. The algorithm focus on the Quality of Service (QoS) experienced by each application (or user). It reduces the waiting and execution times of each individual workflow, unlike other algorithms that give privilege to average completion time of all workflows. The simulation results show that the proposed approach significantly outperforms the other algorithms in terms of individual response time.

This paper addresses the problem of scheduling parallel tasks, represented by a direct acyclic graph (DAG) on heterogeneous clusters. Parallel tasks, also called malleable tasks, are tasks that can be executed on any number of processors with its execution time being a function of the number of processors allotted to it. The scheduling of independent parallel tasks on homogeneous machines has been extensively studied and the case of parallel tasks with precedence constraints has been studied for tree-like graphs. For arbitrary precedence graphs and for heterogeneous machines, the optimization problem is more complex because the processing time of a given task depends on the number of processors and on the total processing capacity of those processors. This paper presents a list scheduling algorithm to minimize the total length of the schedule (makespan) of a given set of parallel tasks, whose dependencies are represented by a DAG