1,721,067 research outputs found
A study on performance measures for auto-scaling CPU-intensive containerized applications
Full text linkAutoscaling of containers can leverage performance measures from the different layers of the computational stack. This paper investigate the problem of selecting the most appropriate performance measure to activate auto-scaling actions aiming at guaranteeing QoS constraints. First, the correlation between absolute and relative usage measures and how a resource allocation decision can be influenced by them is analyzed in different workload scenarios. Absolute and relative measures could assume quite different values. The former account for the actual utilization of resources in the host system, while the latter account for the share that each container has of the resources used. Then, the performance of a variant of Kubernetes’ auto-scaling algorithm, that transparently uses the absolute usage measures to scale-in/out containers, is evaluated through a wide set of experiments. Finally, a detailed analysis of the state-of-the-art is presented
Autonomic orchestration of containers: Problem definition and research challenges
No full textToday, a new technology is going to change the way cloud platforms are designed and managed. This technology is called container. A container is a software environment where to install an application or application component and all the library dependencies, the binaries, and a basic configuration needed to run the application. The container technology promises to solve many cloud application issues, for example the application portability problem and the virtual machine performance overhead problem. The cloud industry is adopting the container technology both for internal usage and as commercial offering. However, we are far away from the maturity stage and there are still many research challenges to be solved. One of them is container orchestration, that make it possible to define how to select, deploy, monitor, and dynamically control the configuration of multi-container packaged applications in the cloud. This paper presents the problem of autonomic container orchestration, analyze the state of the art solutions and discuss open challenge
Dependencies discovery and analysis in distributed systems (short paper)
No full textThe welfare of our daily life depends, even more, on the correct functioning of complex distributed applications. Moreover, new paradigms such as Service oriented computing and Cloud computing encourage the design of application realized coupling services running on different nodes of the same data center or distributed in a geographic fashion. Dependencies discovery and analysis (DDA) is core for the identification of critical and strategical assets an application depends on, and it is valid support to risk and impact analysis [10
An Autonomic Legal-Rule Aware Cloud Service Broker
No full textThe ICT industry, and specifically critical sectors such as healthcare, transportation, energy and government require as mandatory the compliance of the ICT systems and services with legislation and regulation, as well as with standards. In the era of cloud computing, and particularly in a public cloud scenario, this law and regulation compliance management issue is exacerbated by the distributed nature of the system and by the limited control of the customer on the infrastructure/services. Also if the cloud industry is aware of this legislation/regulation compliance issue (e.g. the compliance program of Amazon, Google and Microsoft Azure), right now, there are no mechanism/architectures capable to check and to assure that the compliance is guaranteed during the whole life cycle of a cloud service, off-line and at run-time. In this paper we outline a reference architecture for the autonomic and legislation-aware cloud service broker and we propose a run-time linear programming based model that consider legal constraints and that perform service adaptation for the assurance of QoS and legal rule compliance
ASiMOV: A self-protecting control application for the smart factory
No full textThe evolution of manufacturing systems into a smart factory brings advantages but also increased cyber-risks. This paper investigates the problem of intrusion detection and autonomous response to cyber-attacks targeting the control logic of industrial control applications for the smart factory. Specifically, we propose ASiMOV (Asynchronous Modular Verification), a self-protecting architecture for cyber–physical systems realizing a verifiable control application. ASiMOV is inspired by modular redundancy and leverages virtualization technologies to respond and to prevent cyber-attacks to the control logic. Using simulation experiments, we evaluate: the effects of an attack on an industrial control application enhanced by ASiMOV; the delay introduced by ASiMOV within a control loop; and the cyber-attack detection delay. Results show that, in the simulated scenario, the controller can work with a sampling rate of up to 200 Hertz. Any tampering with the control logic is detected without false positives/negatives in a time equal to the latency between the proposed control application and the proposed IDS (e.g., tens to hundreds of milliseconds)
Decentralized Task Scheduling in Satellite Edge Computing
No full textSatellite Edge Computing has been recently introduced to deploy innovative computational services in space using Low Earth Orbit (LEO) satellite constellations as a distributed computational platform. Running a distributed computing platform in space introduces new challenges to traditional problems like computation offloading, task scheduling, mobility management, fault detection, and recovery. This research focuses on the problem of task scheduling, proposing a system model that accounts for the dynamics of the Satellite Edge Computing environment and a formulation of the scheduling problem as an optimization problem that minimizes the average task response time under constraints on available resources and task completion deadlines. Then, we propose a decentralized algorithm that estimates the task response time and computes a scheduling solution in a fixed time, which depends only on the number of Inter Satellite Links a satellite has (typically four). Finally, we estimate and compare the overhead of the decentralized versus the decentralized solutions, showing the advantages of the proposed approach. Simulation experiments allow us to compare the performance of the decentralized approach with the performance of baseline decentralized and centralized solutions. Results show that, in all scenarios considered, the proposed decentralized algorithm performs better than the baseline centralized and decentralized solutions and is more scalable and highly available
Mechanisms for SLA provisioning in cloud-based service providers
No full textA challenge in cloud resource management is to design self-adaptable solutions capable to react to unpredictable workload fluctuations and changing utility principles. This paper analyzes the problem from the perspective of an Application Service Provider (ASP) that uses a cloud infrastructure to achieve scalable provisioning of its services in the respect of QoS constraints. First we draw a taxonomy of IaaS provider and use the identified features to drive the design of four autonomic service management architectures differing on the degree of control an ASP have on the system. We implemented two of this solutions and related mechanism to test five different resource provisioning policies. The implemented testbed has been evaluated under a realistic workload based on Wikipedia access traces on Amazon EC2 platform. The experimental evaluation performed confirms that: the proposed policies are capable to properly dimension the system resources making the whole system self-adaptable respect to the workload fluctuation. Moreover, having full control over the resource management plan allow to save up to the 32% of resource allocation cost always in the respect of SLA constraints
QoS in grid computing
No full textLarge-scale grids are complex systems composed of thousands of components from disjoined domains. In such environments, planning the capacity to guarantee quality of service (QoS) is a challenge because global service-level agreements (SLAs) depend on local SLAs. A motivating example for grid computing in an enterprise environment is presented. Specifically by using an example of a large insurance company (IC) that offers coverage for vehicles, boats, homes, and businesses, it is shown how resource allocation affects SLA
Auto-scaling of containers: the impact of relative and absolute metrics
No full textToday, The cloud industry is adopting the container technology both for internal usage and as commercial offering. The use of containers as base technology for large-scale systems opens many challenges in the area of resource management at run-Time. This paper addresses the problem of selecting the more appropriate performance metrics to activate auto-scaling actions. Specifically, we investigate the use of relative and absolute metrics. Results demonstrate that, for CPU intense workload, the use of absolute metrics enables more accurate scaling decisions. We propose and evaluate the performance of a new autoscaling algorithm that could reduce the response time of a factor between 0.66 and 0.5 compared to the actual Kubernetes' horizontal auto-scaling algorithm
Measuring Docker Performance : What a Mess!!!
No full textToday, a new technology is going to change the way platforms for the internet of services are designed and managed. This technology is called container (e.g. Docker and LXC). The internet of service industry is adopting the container technology both for internal usage and as commercial offering. The use of container as base technology for large-scale systems opens many challenges in the area of resource management at run-time, for example: autoscaling, optimal deployment and monitoring. Specifically, monitoring of container based systems is at the ground of any resource management solution, and it is the focus of this work. This paper explores the tools available to measure the performance of Docker from the perspective of the host operating system and of the virtualization environment, and it provides a characterization of the CPU and disk I/O overhead introduced by containers
- …
