Fine-grained Resource Management Research Articles

Revamping Cloud Gaming With Distributed Engines

While cloud gaming has brought considerable advantages for its customers, from the point of view of cloud providers, multiple aspects related to infrastructure management still fall short of such kind of service. Indeed, differently from traditional cloud-ready applications, modern game engines are still based on monolithic software architectures. This aspect precludes the applicability of fine-grained resource management and service orchestration schemes, ultimately leading to poor cost-effectiveness. To mitigate these shortcomings, we propose a Cloud-Oriented Distributed Engine for Gaming (CODEG). Thanks to its distributed nature, CODEG is capable of fully exploiting the resource heterogeneity present in cloud data centers, while providing the possibility of spanning its service on multiple network layers up to the edge clouds.

Online Microservice Orchestration for IoT via Multiobjective Deep Reinforcement Learning

By providing loosely coupled, lightweight, and independent services, the microservice architecture is promising for large-scale and complex service provision requirements in the Internet of Things (IoT). However, it requires more fine-grained resource management and orchestration for service provision. Most of the existing microservice orchestration solutions are based on those designed for the traditional cloud. They can only provide coarse-grained resource allocation using possibly conflicting weighted objectives. In this article, we present a fine-grained microservice orchestration approach to provide services online for dynamic requests of IoT applications. By using a fine-grained resource model of energy cost and service end-to-end response time of orchestrated microservices, we formulate the microservice orchestration problem as a multiobjective Markov decision process. We then propose a multiobjective optimization solution based on deep reinforcement learning (DRL) to simultaneously reduce energy consumption and response time. Through extensive experiments, our proposed algorithm presents significant performance results than the state of the art. To the best of our knowledge, this is the first work that addresses microservice orchestration using DRL for multiple conflicting objectives.

ComFlex: Composable and Flexible Resource Management for the IoT

The Internet of Things (IoT) enables new services, such as smart home and smart healthcare integrating various resources of networked devices. However, managing IoT resources is very challenging because a device may have multiple different resources and the service accesses subsets of the resources in multiple devices. To simplify the resource management, the existing IoT frameworks provide resource models and protocols for manufacturers and programmers, but their access granularity is too coarse grained causing overprivilege problems, or too fine grained causing management overheads. To avoid the overprivilege problems without the management overheads, this work proposes a new composable and flexible resource management scheme and implements its prototype compiler-runtime framework called ComFlex. The ComFlex compiler allows manufacturers to register their devices as composition of fine-grained resources by inheriting the existing interfaces of the resources, and programmers to define their own access granularity as a virtual resource that consists of the fine-grained resources of multiple devices. The ComFlex runtime supports fine-grained access control without additional overheads by mapping the fine-grained resources of the virtual resource into physical resources of different devices. To evaluate the ComFlex framework, this work implements 52 resources on 14 devices and six IoT services with ComFlex and an existing fine-grained resource management scheme. Compared to the existing scheme, the ComFlex framework supports the IoT services with 40.2% fewer lines of code and 80.9% less discovery time, without any overprivilege problem and response time delay.

New Scheduling Algorithms for Improving Performance and Resource Utilization in Hadoop YARN Clusters

The MapReduce framework has become the defacto scheme for scalable semi-structured and un-structured data processing in recent years. The Hadoop ecosystem has evolved into its second generation, Hadoop YARN, which adopts fine-grained resource management schemes for job scheduling. Nowadays, fairness and efficiency are two main concerns in YARN resource management because resources in YARN are shared and contended by multiple applications. However, the current scheduling in YARN does not yield the optimal resource arrangement, unnecessarily causing idle resources and inefficient scheduling. It omits the dependency between tasks which is extremely crucial for the efficiency of resource utilization as well as heterogeneous job features in real application environments. We thus propose a new YARN scheduler which can effectively reduce the makespan (i.e., the total execution time) of a batch of MapReduce jobs in Hadoop YARN clusters by leveraging the information of requested resources, resource capacities and dependency between tasks. For accommodating heterogeneity in MapReduce jobs, we also extend our scheduler by further considering the job iteration information in the scheduling decisions. We implemented the new scheduling algorithm as a pluggable scheduler in YARN and evaluated it with a set of classic MapReduce benchmarks. The experimental results demonstrate that our YARN scheduler effectively reduces the makespans and improves resource utilizations.

Minimizing Financial Cost of DDoS Attack Defense in Clouds With Fine-Grained Resource Management

As the cloud systems gain in popularity, they suffer from cyber attacks. One of the most notorious cyber attacks is Distributed Denial of Service (DDoS) attack, which aims to drain the system resources so that the system becomes unresponsive to the genuine users. DDoS attack and defense essentially revolve around resource competition. Many efforts have been made from the perspective of resource investment and management. However, these defending schemes assume that the resources available to defend the attacks are unlimited without taking the financial cost into account. Such coarse-grained defense strategies could cause the problem of resource overprovisioning, which would incur unwanted extra costs to the defender. To tackle this issue, we systematically investigate the problem and propose a birth-death-based fine-grained resource management mechanism, which can both scale in/out and scale down/up. That is, the proposed mechanism adaptively selects the optimal resource leasing mode for cloud service customers so that they can defeat the DDoS attack with minimal financial cost. Extensive analyses and empirical data-based experiments are conducted. The results show both the effectiveness and efficiency of the proposed approach. Comparing to existing work, our proposal can averagely save 53.58% (up to 93.75%) of the cost for the attack defense.

Be Good Neighbors: A Novel Application Isolation Metric Used to Optimize the Initial Container Placement in CaaS

Many cloud providers benefit from fine-grained resource management in Container as a Service (CaaS), e.g., energy savings. However, while optimizing power consumption, cloud providers also need to consider service performance. Few studies in the literature performed a balanced optimization of the two objectives in the initial container placement (ICP). As such, the obtained solutions are often locally optimal. In this work, we model the ICP as a bi-objective optimization problem. Unfortunately, directly applying a conventional performance modeling method in solving the problem is complex and costly, especially for a large-scale application scenario. It also fails to capture the impacts of some non-resource constraints such as high availability constraints. To tackle these issues, we propose a novel application isolation metric to quantify the overall service performance. With regard to the power consumption objective, we opt for a non-linear model since a linear one makes power consumption indistinguishable in a homogeneous setting. With the two building blocks, we introduce an optimization algorithm named First Fit based improving Genetic Algorithm (denoted by FF-based-IGA) to find the best ICP solution. Our experimental results are as follows: 1) the proposed application isolation metric is effective at improving service performance, and 2) FF-based-IGA achieves a desirable balance between power consumption and service performance.

CPPM: a lightweight performance prediction middleware for cloud platforms

As more and more commercial clouds have been applied in various areas, how to evaluate the performance of a cloud platform has become an important issue that needs to be addressed. Furthermore, an effective performance prediction mechanism is of significant value for improving the current cloud services, such as resource allocation and task scheduling. In the paper, we present the design and prototype implementation of a performance prediction system, namely cloud performance prediction middleware (CPPM), which is aiming at providing a set of lightweight and flexible services on existing cloud infrastructure so as to allow cloud providers monitoring, estimating and predicting the runtime performance from various aspects. The CPPM enables cloud providers to make more efficient and fine-grained resource management and scheduling policies based on their short-term workload prediction mechanism; also it provides an application-level performance prediction service which uses skeleton approach to capture execution characteristics of the running applications so as to predict their actual runtime performance and efficiency. Extensive experiments are conducted to examine the effectiveness and efficiency of the CPPM.

CPPM: a lightweight performance prediction middleware for cloud platforms

As more and more commercial clouds have been applied in various areas, how to evaluate the performance of a cloud platform has become an important issue that needs to be addressed. Furthermore, an effective performance prediction mechanism is of significant value for improving the current cloud services, such as resource allocation and task scheduling. In the paper, we present the design and prototype implementation of a performance prediction system, namely cloud performance prediction middleware (CPPM), which is aiming at providing a set of lightweight and flexible services on existing cloud infrastructure so as to allow cloud providers monitoring, estimating and predicting the runtime performance from various aspects. The CPPM enables cloud providers to make more efficient and fine-grained resource management and scheduling policies based on their short-term workload prediction mechanism; also it provides an application-level performance prediction service which uses skeleton approach to capture execution characteristics of the running applications so as to predict their actual runtime performance and efficiency. Extensive experiments are conducted to examine the effectiveness and efficiency of the CPPM.

Efficient jobs scheduling approach for big data applications

Abstract The MapReduce framework has become a leading scheme for processing large-scale data applications in recent years. However, big data applications executed on computer clusters require a large amount of energy, which costs a considerable fraction of the data center’s overall costs. Therefore, for a data center, how to reduce the energy consumption becomes a critical issue. Although Hadoop YARN adopts fine-grained resource management schemes for job scheduling, it doesn’t consider the energy saving problem. In this paper, an Energy-aware Fair Scheduling framework based on YARN (denoted as EFS) is proposed, which can effectively reduce energy consumption while meet the required Service Level Agreements (SLAs). EFS not only can schedule jobs to energy-efficiency nodes, but also can power on or off the nodes. To do so, the energy-aware dynamic capacity management with deadline-driven policy is used to allocate the resources for MapReduce tasks in terms of the average execution time of containers and users request resources. And then, Energy-aware fair based scheduling problem is modeled as multi-dimensional knapsack problem (MKP) and the energy-aware greedy algorithm (EAGA) is proposed to realize tasks fine-grained placement on energy-efficient nodes. Finally, the nodes which have been kept in idle state for the threshold duration are turned off to reduce energy costs. We perform extensive experiments on the Hadoop YARN clusters to compare the energy consumption and executing time of EFS with some state-of-the-art policies. The experimental results show that EFS can not only keep the proper number of nodes in on states to meet the computing requirements but also achieve the goal of energy savings.

Solving a wind turbine maintenance scheduling problem

Driven by climate change mitigation efforts, the wind energy industry has significantly increased in recent years. In this context, it is essential to make its exploitation cost-effective. Maintenance of wind turbines therefore plays an essential role in reducing breakdowns and ensuring high productivity levels. In this paper, we discuss a challenging maintenance scheduling problem rising in the onshore wind power industry. While the research in the field primarily focuses on condition-based maintenance strategies, we aim to address the problem on a short-term horizon considering the wind speed forecast and a fine-grained resource management. The objective is to find a maintenance plan that maximizes the revenue from the electricity production of the turbines while taking into account multiple task execution modes and task-technician assignment constraints. To solve this problem, we propose a constraint programming-based large neighborhood search (CPLNS) approach. We also propose two integer linear programming formulations that we solve using a commercial solver. We report results on randomly generated instances built with input from wind forecasting and maintenance scheduling software companies. The CPLNS shows an average gap of 1.2% with respect to the optimal solutions if known, or to the best upper bounds otherwise. These computational results demonstrate the overall efficiency of the proposed metaheuristic.

Multi-Tier Resource Allocation for Data-Intensive Computing

As distributed computing systems are used more widely, driven by trends such as 'big data' and cloud computing, they are being used for an increasingly wide range of applications. With this massive increase in application heterogeneity, the ability to have a general purpose resource management technique that performs well in heterogeneous environments is becoming increasingly important.In this paper, we present Multi-Tier Resource Allocation (MTRA) as a novel fine-grained resource management technique for distributed systems. The core idea is based on allocating resources to individual tasks in a tiered or layered approach. To account for heterogeneity, we propose a dynamic resource allocation method that adjusts resource allocations to individual tasks on a cluster node based on resource utilisation levels. We demonstrate the efficacy of this technique in a data-intensive computing environment, MapReduce data processing framework in Hadoop YARN. Our results demonstrate that MTRA is an effective general purpose resource management technique particularly for data-intensive computing environments. On a range of MapReduce benchmarks in a Hadoop YARN environment, our MTRA technique improves performance by up to 18%. In a Facebook workload model it improves job execution times by 10% on average, and up to 56% for individual jobs.