Virtualized Data Centers Research Articles

Energy-Aware Scheduling for Precedence-Constrained Parallel Virtual Machines in Virtualized Data Centers

Large scale Internet services are expected to only increase in complexity and popularity. Their energy consumption is also a major concern in data centers. Smart scheduling of their sub-services on data center Physical Machines (PM) can effectively improve their energy as well as performance. Since today servers are not energy-proportional yet, a major and traditionally neglected source of inefficiency in them is the utilization level of PMs. We present two scheduling algorithms for precedence-constrained parallel Virtual Machines (VM) in a virtualized data center where each VM represents a sub-service of the Internet-scale service. Our algorithms use virtualization technology to increase utilization of the PMs, and hence reduce total number of active PMs, to improve energy with minimal effect on makespan. Both proposed algorithms have a polynomial time complexity which make them suitable options for scheduling of large services. Simulation results using real-world services demonstrate that the algorithms are capable of increasing utilization level of PMs on average by 52 % and improving energy consumption by 18 % while the makespan of services is degraded less than 2 %.

Profiling-Based Workload Consolidation and Migration in Virtualized Data Centers

Improving energy efficiency of data centers has become increasingly important nowadays due to the significant amounts of power needed to operate these centers. An important method for achieving energy efficiency is server consolidation supported by virtualization. However, server consolidation may incur significant degradation to workload performance due to virtual machine (VM) co-location and migration. How to reduce such performance degradation becomes a critical issue to address. In this paper, we propose a profiling-based server consolidation framework which minimizes the number of physical machines (PMs) used in data centers while maintaining satisfactory performance of various workloads. Inside this framework, we first profile the performance losses of various workloads under two situations: running in co-location and experiencing migrations. We then design two modules: (1) consolidation planning module which, given a set of workloads, minimizes the number of PMs by an integer programming model, and (2) migration planning module which, given a source VM placement scenario and a target VM placement scenario, minimizes the number of VM migrations by a polynomial time algorithm. Also, based on the workload performance profiles, both modules can guarantee the performance losses of various workloads below configurable thresholds. Our experiments for workload profiling are conducted with real data center workloads and our experiments on our two modules validate the integer programming model and the polynomial time algorithm.

Development of collisional data base for elementary processes of electron scattering by atoms and molecules

We present a progress report on the development of the Belgrade electron/molecule data base which is hosted by The Institute of Physics, University of Belgrade and The Astronomical Observatory Belgrade. The data base has been developed under the standards of Virtual Atomic Molecular Data Centre (VAMDC) project which provides a common portal for several European data bases that maintain atomic and molecular data. The Belgrade data base (BEAMDB) covers collisional data of electron interactions with atoms and molecules in the form of differential (DCS) and integrated cross sections as well as energy loss spectra. The final goal of BEAMDB becoming both a node within the VAMDC consortium and within the radiation damage RADAM data base has been achieved.

SPA: Smart Placement Approach for Cloud-service Datacenter Networks

Abstract Business disciplines expand rapidly, so does the tendency to rely on computer applications and its varying services to support such expansion. Often, this is achieved through introducing physical network infrastructures that provide the appropriate environ- ments to run such applications. The required services change rapidly, and accordingly its resource requirements. In most cases, this may require building new physical networks which could lead to low utilization rates and high service-costs. A promising ap- proach that is becoming increasingly popular to overcome such a problem is known as Virtual Datacenter Networks (VDNs). These VDNs are usually hosted over physical networks; overlaying its resources to gain the dynamic required services. Cloud-service Datacenter Networks (CDNs) can provide such a service under a delivery model that is called Infrastructure as a Service (IaaS). In this paper, we present a Smart Placement Approach (SPA) that provides for smart placement maps of VDNs over CDNs. In this, we point out that choosing the placement maps for such VDNs should satisfy its requirements while: maintaining load-balancing over the hosting CDNs, guaranteeing its Quality of Service (QoS) levels, and assuring low placement costs.

Efficient Virtual Machine Placement with Energy Saving in Cloud Data Center

Cloud data centers provide computing infrastructure as a service to their customers on pay per use basis. In virtualized data centers CPU, RAM, storage and bandwidth are allotted to a Virtual Machine (VM) from pool of shared resources. An autonomic consolidation of VMs on appropriate Physical Machine (PM) by achieving performance and saving cost is the key challenge for virtualized data centers. This paper presents a self-organizing and multi-objective approach for autonomic consolidation of VMs. The proposed approach does the initial placement of VMs in appropriate PM of cloud data centre which addresses different issues altogether such as maximum resource requirement during setup of VMs, future demand of free resources at peak load, improving the performance and energy saving by keeping idle PMs at offline state. The performance of the proposed algorithm is evaluated by simulating a data center with randomly generated resource capacities of PMs and resource requirement of VMs. Experiment results of proposed technique are also compared with standard algorithms of VM consolidation such as first-first, next-fit and random-selection on two dimensions of resources-CPU and RAM.

A solution of dynamic VMs placement problem for energy consumption optimization based on evolutionary game theory

Power saving of data centers has become an urgent problem in recent years. For a virtualized data center, optimizing the placement of virtual machines (VMs) dynamically is one of the most effective methods for power savings. Based on a deep study on VMs placement, a solution is proposed and described in this paper to solve the problem of dynamic placement of VMs toward optimization of their energy consumptions. A computational model of energy consumption is proposed and built. A novel algorithm based on evolutionary game theory is also presented, which successfully addresses the challenges faced by dynamic placement of VMs. It is proved that the proposed algorithm can reach the optimal solutions theoretically. Experimental results also demonstrate that, by adjusting VMs placement dynamically, the energy consumption can be reduced correspondingly. In comparison with the existing state of the arts, our proposed method outperforms other five algorithms tested and achieves savings of 30–40% on energy consumption.

SUP@VAMDC*

Atomic and molecular (A+M) data are critical for a range of applications such as astrophysics, atmospheric physics, fusion, environmental sciences, combustion chemistry, health and clinical science including radiotherapy. They underpin industries ranging from technological plasmas to lighting. The Virtual Atomic and Molecular Data Centre (VAMDC) is a major new European initiative that has built a unified, secure, documented, flexible and interoperable e-science environment-based interface to 24 existing A+M databases. SUP@VAMDC will coordinate collaborative studies and provide the supporting procedures that will develop a global e-infrastructure for the assembly, curation and access to A+M data. It will adopt and adapt VAMDC protocols to include further A+M databases and networks pathways by which common policies may be developed for an open science version of VAMDC that will extend its availability to new audiences such as schools, students, citizens and industry.

Integration of CO2 spectral line parameters from the CDSD databanks into the virtual atomic and molecular data center (VAMDC)

The spectroscopic banks CDSD-296, CDSD-1000, and CDSD-4000 are presented in the context of integration into the Virtual Atomic and Molecular Data Centre (VAMDC) in order to extend the range of their users. A brief review of the VAMDC system is given concerning the purpose of this project, its infrastructure, and the data representation format. The relational structure of CDSD databases, adapted to VAMDC needs, is presented. Some of the technical problems connected with the huge volumes of stored information are considered.

Cost-Benefit Analysis of Datacenter Consolidation Using Virtualization

The objectives of datacenter consolidation are cost reduction and superior services. A datacenter consolidation plan includes minimizing investments in IT infrastructure and buildings and reducing power consumption related to cooling. Such a process requires scalable planning and implementation. Virtualization is the most popular and cost-effective technology for datacenter consolidation. In this article, the author runs a cost-benefit analysis of virtualization and datacenter consolidation using the Global Virtual datacenter online calculator and VMware's ROI CO ((for) return on investment/total cost of ownership) calculator version 3.0.

Entropy Variation and J48 Algorithm based Intrusion Detection System for Cloud Computing

Cloud Computing has achieved formidable impetus where IT infrastructures and applications are provided as service to end users. It provides shared pool of resources in addition with Data storage, computer processing power and specialized corporate and user applications. Users can access Cloud services any time anywhere and store large amount of data from anywhere, due to increases the popularity of Cloud computing there is risk of Security. Data which is stored on cloud may be vulnerable which is easy for attackers to compromise the virtual machines as zombies and explore these vulnerabilities in cloud system. Because when we move data or information in cloud we do not have any control on that data which can be handling by third party. Hence, there is the vital requirement of more security measures to protect cloud. In this paper we propose an Intrusion detection system which is based on Entropy variation and J48 Decision tree algorithm through which we can detect or prevent vulnerable virtual machines, Data center and Host from being compromised in the cloud also we can protect data and applications in Cloud like wide area network traffic. This proposed solution results gives more accuracy for attack detection and low false alarm rate. For simulation we use Cloud Sim (version 3.1) and used KDDCUP '99 Dataset to evaluate rules and testing datasets to detect intrusion.

Capacity Management for Virtualized Data Centers using ECIES and Scheduling

the growth Internet development cloud computing is novel technique to serve better and secure services. E- business is growing rapidly with the development of Internet. The cloud computing provides on demand self service methodology that authorizes to request resources dynamically as a best benefit. The use of Cloud Computing is ahead reputation due to its mobility and massive availability in minimum cost. Here in this paper an efficient Capacity Management of users data on datacenters is proposed using attribute and scheduling techniques. The proposed technique provides much efficient use of Virtualized data as compared to the existing technique.

Research of Network Virtualization in Data Center

With the rapid grow of the volume of data and internet application, as an efficient and promising infrastructure, data center has been widely deployed .data center provide a variety of perform for network services, applications such as video stream, cloud compute and so on. All this services and applications call for volume, compute, bandwidth, and latency. Existing data centers lacks enough flexible so they provide poor support in QOS, deployability, manageability, and defense when facing attacks. Virtualized data centers are a good solution to these problems. Compared to existing data centers, virtualized data centers do better in resource utilization, scalability, and flexibility.

Fault tolerance mechanisms for virtual data center architectures

A virtual data center (VDC) is a combination of interconnected virtual servers hosted on a physical data center that hosts multiple such VDCs. This enables efficient sharing of the data center's resources while handling dynamic resource requirements of the clients. The SecondNet architecture (Guo et al. in Proceedings of ACMSIGCOMM conference on data communication, Barcelona, pp 63---74, 2009) realizes this VDC concept and includes a centralized VDC resource-mapping (virtual to physical) algorithm. Fault tolerance is an important requirement in data center-based services, in order to increase reliability and availability. In this paper, we propose a fault tolerance mechanism to handle server failures by efficiently migrating the virtual machines (VMs) hosted on the failed server to a new location. Using our mechanism, it is shown that recovery from all the faults is possible, even for a server utilization of 90 %. In order to reduce the impact of server failures on the VDCs hosted in the data center, we then present a new load balancing scheme based on clustering that efficiently allocates the VDCs on the data center. Using this scheme, we were able to reduce the affected number of VMs per server failure by 63 %, in case of a BCube network of size 625 nodes, and by 86 %, in case of a BCube network of size 1,296 nodes.

Improving MapReduce performance by balancing skewed loads

MapReduce has emerged as a popular computing model used in datacenters to process large amount of datasets. In the map phase, hash partitioning is employed to distribute data that sharing the same key across data center-scale cluster nodes. However, we observe that this approach can lead to uneven data distribution, which can result in skewed loads among reduce tasks, thus hamper performance of MapReduce systems. Moreover, worker nodes in MapReduce systems may differ in computing capability due to (1) multiple generations of hardware in non-virtualized data centers, or (2) co-location of virtual machines in virtualized data centers. The heterogeneity among cluster nodes exacerbates the negative effects of uneven data distribution. To improve MapReduce performance in heterogeneous clusters, we propose a novel load balancing approach in the reduce phase. This approach consists of two components: (1) performance prediction for reducers that run on heterogeneous nodes based on support vector machines models, and (2) heterogeneity-aware partitioning (HAP), which balances skewed data for reduce tasks. We implement this approach as a plug-in in current MapReduce system. Experimental results demonstrate that our proposed approach distributes work evenly among reduce tasks, and improves MapReduce performance with little overhead.

Traffic-aware VDC embedding in data center: A case study of fattree

Virtualization is a common technology for resource sharing in data center. To make efficient use of data center resources, the key challenge is to map customer demands (modeled as virtual data center, VDC) to the physical data center effectively. In this paper, we focus on this problem. Distinct with previous works, our study of VDC embedding problem is under the assumption that switch resource is the bottleneck of data center networks (DCNs). To this end, we not only propose relative cost to evaluate embedding strategy, decouple embedding problem into VM placement with marginal resource assignment and virtual link mapping with decided source-destination based on the property of fat-tree, but also design the traffic aware embedding algorithm (TAE) and first fit virtual link mapping (FFLM) to map virtual data center requests to a physical data center. Simulation results show that TAE+FFLM could increase acceptance rate and reduce network cost (about 49% in the case) at the same time. The traffic aware embedding algorithm reduces the load of core-link traffic and brings the optimization opportunity for data center network energy conservation.

Novel resource allocation algorithms to performance and energy efficiency in cloud computing

The rapid growth in demand for computational power has led to a shift to the cloud computing model established by large-scale virtualized data centers. Such data centers consume enormous amounts of electrical energy. Cloud providers must ensure that their service delivery is flexible to meet various consumer requirements. However, to support green computing, cloud providers also need to minimize the cloud infrastructure energy consumption while conducting the service delivery. In this paper, for cloud environments, a novel QoS-aware VMs consolidation approach is proposed that adopts a method based on resource utilization history of virtual machines. Proposed algorithms have been implemented and evaluated using CloudSim simulator. Simulation results show improvement in QoS metrics and energy consumption as well as demonstrate that there is a trade-off between energy consumption and quality of service in the cloud environment.

Using 3-D Virtual Worlds as a Platform for an Experiental Case Study in Information Systems Auditing

The paper is a case study of an experiential exercise in information systems auditing of a virtual data center constructed in the 3-D virtual world of Second Life. A techonology and data center was created and populated with numerous risk exposure for an organization.

Automated and Agile Server ParameterTuning by Coordinated Learning and Control

Automated server parameter tuning is crucial to performance and availability of Internet applications hosted in cloud environments. It is challenging due to high dynamics and burstiness of workloads, multi-tier service architecture, and virtualized server infrastructure. In this paper, we investigate automated and agile server parameter tuning for maximizing effective throughput of multi-tier Internet applications. A recent study proposed a reinforcement learning based server parameter tuning approach for minimizing average response time of multi-tier applications. Reinforcement learning is a decision making process determining the parameter tuning direction based on trial-and-error, instead of quantitative values for agile parameter tuning. It relies on a predefined adjustment value for each tuning action. However it is nontrivial or even infeasible to find an optimal value under highly dynamic and bursty workloads. We design a neural fuzzy control based approach that combines the strengths of fast online learning and self-adaptiveness of neural networks and fuzzy control. Due to the model independence, it is robust to highly dynamic and bursty workloads. It is agile in server parameter tuning due to its quantitative control outputs. We implemented the new approach on a testbed of virtualized data center hosting RUBiS and WikiBench benchmark applications. Experimental results demonstrate that the new approach significantly outperforms the reinforcement learning based approach for both improving effective system throughput and minimizing average response time.

Distributed temperature-aware resource management in virtualized data center

While resource consolidation enables energy efficiency in virtualized data centers, it results in increased power density and causes excessive heat generation. To prevent servers from overheating and avoid potential damage and/or service outages, data centers need to incorporate temperature awareness in resource provisioning decisions. Moreover, data centers are subject to various peak power constraints (such as peak server power) that have to be satisfied at all times for reliability concerns. In this paper, we propose a novel resource management algorithm, called DREAM (Distributed REsource mAnagement with teMperature constraint), to optimally control the server capacity provisioning (via power adjustment), virtual machine (VM) CPU allocation and load distribution for minimizing the data center power consumption while satisfying the Quality of Service (QoS), IT peak power and maximum server temperature constraints. By using DREAM, each server can autonomously adjust its discrete processing speed (and hence, power consumption, too), and optimally decide the VM CPU allocation as well as amount of workloads to process in the hosted VMs, in order to minimize the total power consumption which incorporates both server power and cooling power. We formally prove that DREAM can yield the minimum power with an arbitrarily high probability while satisfying the peak power and server temperature constraints. To complement the analysis, we perform a simulation study and show that DREAM can significantly reduce the power consumption compared to the optimal temperature-unaware algorithm (by up to 33%) and equal load distribution (by up to 86%).

A TRILL-based multi-tenant data center network

To achieve an efficient ‘Infrastructure as a Service’ (IaaS) service, virtualized data centers have to provide large-scale multi-tenancy services and to meet new requirements in terms of efficiency and availability. The main limitation of current network architectures mainly comes from inadequate design standards, such as those for Ethernet Bridging (802.1D) and Virtual LAN (802.1Q), which were not designed to handle large numbers of tenants and the resulting traffic volumes.In this paper, we present a TRILL based multi-tenant network architecture for virtualized data center networks. Through the TRILL protocol, this architecture achieves the best of both layers (bridging and routing), thus building a large-scale layer-2 network while ensuring scalability, efficiency, fault-tolerance and simplified management. In addition, this architecture integrates the VNT (Virtual Network over TRILL) solution, enabling the creation of large numbers of logical networks within a virtualized data center. The complete implementation of our solution is currently deployed and running in the GANDI public cloud infrastructure.Experimental evaluations show that VNT is able to support several hundreds of switches, thousands of physical nodes, and hundreds of thousands of tenants with only negligible overhead, all with very minimal configuration which may be completely automated. VNT additionally permits backward compatibility with both traditional layer-2 equipment as well as newer hardware supporting TRILL.