MAGNETIC: Multi-Agent Machine Learning-Based Approach for Energy Efficient Dynamic Consolidation in Data Centers

Improving the energy efficiency of data centers while guaranteeing Quality of Service (QoS), together with detecting performance variability of servers caused by either hardware or software failures, are two of the major challenges for efficient resource management of large-scale cloud infrastructures. Previous works in the area of dynamic Virtual Machine (VM) consolidation are mostly focused on addressing the energy challenge, but fall short in proposing comprehensive, scalable, and low-overhead approaches that jointly tackle energy efficiency and performance variability. Moreover, they usually assume over-simplistic power models, and fail to accurately consider all the delay and power costs associated with VM migration and host power mode transition. These assumptions are no longer valid in modern servers executing heterogeneous workloads and lead to unrealistic or inefficient results. In this paper, we propose a centralized-distributed low-overhead failure-aware dynamic VM consolidation strategy to minimize energy consumption in large-scale data centers. Our approach selects the most adequate power mode and frequency of each host during runtime using a distributed multi-agent Machine Learning (ML) based strategy, and migrates the VMs accordingly using a centralized heuristic. Our Multi-AGent machine learNing-based approach for Energy efficienT dynamIc Consolidation (MAGNETIC) is implemented in a modified version of the CloudSim simulator, and considers the energy and delay overheads associated with host power mode transition and VM migration, and is evaluated using power traces collected from various workloads running in real servers and resource utilization logs from cloud data center infrastructures. Results show how our strategy reduces data center energy consumption by up to 15% compared to other works in the state-of-the-art (SoA), guaranteeing the same QoS and reducing the number of VM migrations and host power mode transitions by up to 86% and 90%, respectively. Moreover, it shows better scalability than all other approaches, taking less than 0.7% time overhead to execute for a data center with 1500 VMs. Finally, our solution is capable of detecting host performance variability due to failures, automatically migrating VMs from failing hosts and draining them from workload.

Integrating Heuristic and Machine-Learning Methods for Efficient Virtual Machine Allocation in Data Centers

David Atienza Alonso, Ali Pahlevan, Xiaoyu Qu, Marina Zapater Sancho

Modern cloud data centers (DCs) need to tackle efficiently the increasing demand for computing resources and address the energy efficiency challenge. Therefore, it is essential to develop resource provisioning policies that are aware of virtual machine (VM) characteristics, such as CPU utilization and data communication, and applicable in dynamic scenarios. Traditional approaches fall short in terms of flexibility and applicability for large-scale DC scenarios. In this paper we propose a heuristic- and a machine learning (ML)-based VM allocation method and compare them in terms of energy, quality of service (QoS), network traffic, migrations, and scalability for various DC scenarios. Then, we present a novel hyper-heuristic algorithm that exploits the benefits of both methods by dynamically finding the best algorithm, according to a user-defined metric. For optimality assessment, we formulate an integer linear programming (ILP)-based VM allocation method to minimize energy consumption and data communication, which obtains optimal results, but is impractical at runtime. Our results demonstrate that the ML approach provides up to 24% server-to-server network traffic improvement and reduces execution time by up to 480x compared to conventional approaches, for large-scale scenarios. On the contrary, the heuristic outperforms the ML method in terms of energy and network traffic for reduced scenarios. We also show that the heuristic and ML approaches have up to 6% energy consumption overhead compared to ILP-based optimal solution. Our hyper-heuristic integrates the strengths of both the heuristic and the ML methods by selecting the best one during runtime.

2017

Multi-Objective System-Level Management of Modern Green Data Centers

Ali Pahlevan

In our modern society, the average citizen has turned into a daily cloud user. Despite being virtually transparent for the user, internet services require data centers behind the scenes. Data centers burn several megawatts of power and their electricity bill is a major fraction of their costs. Hence, modern cloud data centers need to tackle efficiently the increasing demand for computing resources while at the same time addressing the energy efficiency challenge. This is a complex optimization problem that worsens as more constraints and objectives are added. In public clouds, virtualization transforms a data center into a flexible cloud infrastructure in which Virtual Machines (VMs) behave as separate entities that share physical hardware resources. Therefore, it is essential to develop resource provisioning policies that are aware of VMs characteristics and applicable in dynamic scenarios. To address these challenges, this thesis first presents heuristic and Machine Learning (ML)-based VM allocation methods for various data center scenarios. Then, a novel hyper-heuristic algorithm is proposed that exploits the benefits of both methods by dynamically finding the best one according to a user-defined metric. However, optimizing the energy and cost of a single data center powered by the grid is not enough in today's cloud computing, where multiple data centers are used to deploy online services, and utilize renewable energy sources to reduce their carbon footprint. This thesis also presents a two-phase multi-objective VM placement along with a dynamic migration technique, for geo-distributed data centers coupled with renewable and Electrical Energy Storage (EES) sources to tackle the challenges of operational cost optimization and energy-performance trade-offs. Furthermore, in order to efficiently minimize cost, power market operators have recently introduced emerging demand-response programs, in which electricity consumers regulate their power usage following providers' requests. Therefore, it is essential to develop bidding strategies for data centers to participate in emerging power markets together with power management policies that are aware of power market requirements at run-time. In this thesis I also propose a strategy to jointly optimize the data center demand-response provision problem and VM allocation that satisfies the hour-ahead power market constraints in the presence of renewable and EES energy. Even if novel data center resource management techniques can efficiently tackle the dramatic increase in the number of servers, each computing server remains power limited due to effect of post-Dennard scaling. Therefore, techniques such as Near-Threshold Computing (NTC) need to complement novel system-level approaches to improve data centers' energy efficiency. For this purpose, I first use an accurate power modeling characterization for a new server architecture based on the Fully Depleted Silicon On Insulator (FD-SOI) process technology that enables NTC features. Then, I explore the new energy-performance trade-offs brought by NTC servers when executing virtualized applications. Finally, based on this analysis, I propose an energy proportionality-aware VM allocation method at data center level that exploits the knowledge of VMs characteristics together with the accurate power model presented for NTC servers. As a result, the proposed approach increases the energy proportionality of next-generation NTC-based data centers.

EPFL2019

Multi-Objective System-Level Management of Modern Green Data Centers

Ali Pahlevan

EPFL2019