Physical Host Research Articles

Virtual Network Functions Routing and Placement for Edge Cloud Latency Minimization

As a new way to design, deploy, and manage network services, network functions virtualization (NFV) decouples the network functions, from one or more physical network infrastructures and black boxes so they can run in software. It therefore comes as no surprise that NFV originated from service providers, who were looking to improve the deployment of new network services to support their revenue and growth objectives. Within the NFV ecosystem, high availability, and low latency are one of the key quality of service (QoS) benefits that service providers can expect from the 5G Cloud and the NFV networks to make delay-critical services such as remote surgery a reality. Therefore, network services should be placed, chained, and routed through the network considering users/tenants stringent QoS and service-level agreement requirements. To this end, routing and placement optimization plays a major role in improving network performance and the overall network cost. In this paper, we study the problem of virtual network functions (VNFs) placement and routing across the physical hosts to minimize overall latency defined as the queuing delay within the edge clouds and in network links. In that respect, this paper takes a holistic view by considering not only VNFs chaining and placement problem but also considering the flows routing aspect since these two problems are inter-related and have a major impact on network latency.

Meta-heuristic based framework for workflow load balancing in cloud environment

Cloud services are based on datacenter which provides resources on demand with higher capacity, lowest response time and improved resource utilization. The data center comprises of physical hosts which are effectively utilized in the form of Virtual Machines. The task scheduling problem is the mapping of tasks to suitable resources (VMs) as required and it is NP-hard problem. Further, the scheduling algorithms are followed by load balancing techniques for efficient utilization of VMs. In this paper a framework for Load balancing in Cloud Environment has been proposed and implemented for overflow and underflow VM identification. Two metaheuristics and one heuristic have been used in the proposed framework to achieve effective and efficient utilization of VMs in cloud environment. Further, the performance of the proposed framework has been analysed on the basis of makespan and cost metrics which are computed while executing scientific workflow tasks.

Uncovering the Security Implications of Cloud Multi-Tenancy with Bolt

Cloud providers routinely schedule multiple applications per physical host to increase efficiency. The resulting interference on shared resources often leads to performance degradation and, more importantly, security vulnerabilities. Interference can leak important information about an application, ranging from a services placement to confidential data, such as private keys. We present Bolt, a practical system that accurately detects the type and characteristics of applications sharing a cloud platform based on the interference an adversary sees on shared resources. Bolt leverages online data mining techniques that only require 2-5 seconds for detection. In a multi-user study on Amazon Elastic Compute Cloud (EC2), Bolt correctly identifies the characteristics of 385 out of a set of 436 diverse workloads. Extracting this information enables a wide spectrum of previously impractical cloud attacks, including denial of service (DoS) attacks that increase tail latency by 140X, as well as resource freeing attacks (RFAs), and co-residency attacks. Finally, we show that, while advanced isolation mechanisms such as cache partitioning lower detection accuracy, they are insufficient to eliminate these vulnerabilities altogether. To do so, one must either disallow core sharing or allow it only between threads of the same application, leading to significant inefficiencies and performance penalties.

Handover: A mechanism to improve the reliability and availability of network services for clients behind a network address translator

With the rapid growth of Infrastructure as a Service (IaaS), it becomes more important to increase reliability and availability of a service. In our patent, called Connector in the paper, we allow a server to be live migrated to another physical host with a different Internet Protocol (IP) address; meanwhile, all existing clients still can use their original connections to connect to the new server. However, a client sitting behind a network address translator cannot use Connector directly. In this research, we propose a new mechanism, called Handover, to allow a client sitting behind a network address translator to use Connector. We apply a fake three-way handshake to prevent the redirected traffic from being blocked by a network address translator router. Experimental results show that Handover is effective and the overhead of this changeover process is less than 0.2 s. Furthermore, it may be integrated into a Distributed Denial of Service (DDoS) defense system to guard a host against DDoS attacks.

Network slicing to improve multicasting in HPC clusters

In high performance computing (HPC) resources’ extensive experiments are frequently executed. HPC resources (e.g. computing machines and switches) should be able to handle running several experiments in parallel. Typically HPC utilizes parallelization in programs, processing and data. The underlying network is seen as the only non-parallelized HPC component (i.e. no dynamic virtual slicing based on HPC jobs). In this scope we present an approach in this paper to utilize software defined networking (SDN) to parallelize HPC clusters among the different running experiments. We propose to accomplish this through two major components: A passive module (network mapper/remapper) to select for each experiment as soon as it starts the least busy resources in the network, and an SDN-HPC active load balancer to perform more complex and intelligent operations. Active load balancer can logically divide the network based on experiments’ host files. The goal is to reduce traffic to unnecessary hosts or ports. An HPC experiment should multicast, rather than broadcast to only cluster nodes that are used by the experiment. We use virtual tenant network modules in Opendaylight controller to create VLANs based on HPC experiments. In each HPC host, virtual interfaces are created to isolate traffic from the different experiments. The traffic between the different physical hosts that belong to the same experiment can be distinguished based on the VLAN ID assigned to each experiment. We evaluate the new approach using several HPC public benchmarks. Results show a significant enhancement in experiments’ performance especially when HPC cluster experiences running several heavy load experiments simultaneously. Results show also that this multi-casting approach can significantly reduce casting overhead that is caused by using a single cast for all resources in the HPC cluster. In comparison with InfiniBand networks that offer interconnect services with low latency and high bandwidth, HPC services based on SDN can provide two distinguished objectives that may not be possible with InfiniBand: The first objective is the integration of HPC with Ethernet enterprise networks and hence expanding HPC usage to much wider domains. The second objective is the ability to enable users and their applications to customize HPC services with different QoS requirements that fit the different needs of those applications and optimize the usage of HPC clusters.

Energy-efficient Tasks Scheduling Heuristics with Multi-constraints in Virtualized Clouds

Reducing energy consumption has become an important task in cloud datacenters. Many existing scheduling approaches in cloud datacenters try to consolidate virtual machines (VMs) to the minimum number of physical hosts and hence minimize the energy consumption. VM live migration technique is used to dynamically consolidate VMs to as few PMs as possible; however, it introduces high migration overhead. Furthermore, the cost factor is usually not taken into account by existing approaches, which will lead to high payment cost for cloud users. In this paper, we aim to achieve energy reduction for cloud providers and payment saving for cloud users, and at the same time, without introducing VM migration overhead and without compromising deadline guarantees for user tasks. Motivated by the fact that some of the tasks have relatively loose deadlines, we can further reduce energy consumption by proactively postponing the tasks without waking up new physical machines (PMs). A heuristic task scheduling algorithm called Energy and Deadline Aware with Non-Migration Scheduling (EDA-NMS) algorithm is proposed, which exploits the looseness of task deadlines and tries to postpone the execution of the tasks that have loose deadlines in order to avoid waking up new PMs. When determining the VM instant types, EDA-NMS selects the instant types that are just sufficient to guarantee task deadline to reduce user payment cost. The results of extensive experiments show that our algorithm performs better than other existing algorithms on achieving energy efficiency without introducing VM migration overhead and without compromising deadline guarantees.

Energy-Aware Virtual Machine Scheduling on Data Centers with Heterogeneous Bandwidths

Ethernet is the main choice of connectivity in data centers operating in clouds. The Ethernet standards body (IEEE 802.3) has generated Ethernet specifications with increasing transmission rates since its inception. As a result, it is observed that data centers with groups of physical hosts having heterogeneous bandwidths. In this paper, we present an energy-aware virtual machine’s scheduling method, taking into account this heterogeneity while aiming at energy efficiency. This method consists of two algorithms: one to determine to which physical hosts virtual machines should be allocated, and the other for the provision of bandwidth on physical hosts to virtual machines. Our study shows that, regarding energy savings for the studied scenarios, the presented method is comparable to other energy-aware methods in data centers with groups of machines with homogeneous settings, surpassing them in groups of machines with heterogeneous configurations, bringing improvements especially to data centers with heterogeneous bandwidths’ networks or to where restrictive SLAs are used for the provision of bandwidth to virtual machines.

A Probabilistic Model for Finding an Optimal Host Framework and Load Distribution in Cloud Environment

Cloud Computing is being widely accepted in this computing world and comprises of many hardware and software resources. It becomes critical to meet their client’s needs on time if the cloud data centers are so overloaded. In this regard, if the utilization of resources becomes an intelligent way so that we focused on the selection problem of the physical hosts for deploying the tasks. So, OPH-LB (Optimal Physical Host with effective Load Balancing) framework is proposed to model the service of client’s requests in this IaaS architecture with heterogeneous virtual machines. It proposes the thought of accomplishing load balancing in this dynamic environment. Firstly, this OPH-LB approach is filtering the qualified hosts among all which accomplishes the requirement of deploying tasks. Then out of those qualified sets, we apply probablistic model which helps to find the most optimal host in terms of its computing capability and its performance function. Further, the performance is analyzed using Cloudsim simulation tool and compared with existing approaches. The results demonstrate that our model has improved the throughput, reduced the failure rate and optimized the attainment of cloud data centers.

Enhancing Energy-Efficient and QoS Dynamic Virtual Machine Consolidation Method in Cloud Environment

Virtual machine (VM) consolidation techniques are a means to improve energy efficiency and the utilization of cloud data center resources. However, aggressive VM consolidation approaches lead to physical host over-utilization and generate massive undesired VM migrations, which cause degradation of the performance of both the hosts and the VM. Additionally, it has been a significant challenge to improve energy efficiency and resource utilization in the data center while delivering services with guaranteed quality of service (QoS). To address the problem, we propose an enhancing energy-efficient and QoS dynamic virtual machine consolidation (EQVC) method, which consists of four algorithms that correspond to different stages in VM consolidation. In this approach, we select redundant VMs from the hosts before they overload and migrate the VMs to other hosts to save energy and guarantee QoS requirements. We also introduce a host-model with adaptive reserved resources to prevent re-overload of hosts. To prove the effectiveness of our proposed method and algorithms, we experiment under different workload traces from a real system. The experimental results demonstrate that EQVC approach can significantly outperform other traditional methods regarding energy consumption, QoS guarantees, and the number of VM migrations.

Performance evaluation of FreeSurfer medical application in cloud computing

FreeSurfer is set of scientific medical applications mostly used to study brain tissue and its related diseases. Its performance is a key point in order to improve the required time to establish a precise diagnostic. In this paper, the suitability of cloud infrastructures to deploy and run FreeSurfer is analyzed. The results show that KVM based cloud infrastructures deployed on commodity hardware can be used to execute this application, since a very low overhead is introduced on its performance when the results are compared with the physical host. FreeSurfer cloud deployments have several advantages as the possibility of scaling the computing resources on demand, as they are required.

PARAMETRIC ANALYSIS TO ENHANCE SECURITY IN CLOUD COMPUTING TO PREVENT ATTACKS IN LIVE MIGRATION

Cloud computing is an emerging field of today’s era to provide numerous services to the users. By using application software’s like net browser, IT users will get these services dynamically over internet. Virtualization technology will offer these varieties of services by live migration. During live migration process VMs are reallocated from one physical host to another without disconnecting the application. Live migration facilitates fault-tolerance, online system maintenance etc. Existing security models does not guarantee complete security of live migration process. So, security is a major issue in LM. In this paper, we tend to discuss the various security concerns, attacks on LM and solution approaches that are Network Isolation (VLAN), v-TPM LM protocol, NSE-H and The CoM Security Framework, role based migration, SPLM Model, The Live Migration Defence Framework and The Protection Aegis for Live Migration (PALM). The comparison among these solutions based on the security dimensions is also given.

Improving security in cloud by formal modeling of IaaS resources

Nowadays, it is a matter of fact that Cloud is a “must” for all complex services requiring great amount of resources. Big-Data Services are a striking example: they actually perform many kind of analysis (like analytics) on very big repositories. Many File Systems and middleware exist for efficient distribution and management of data and they usually use Cloud Resources. Anyway Several problems arose about Security of data: Virtualization is the base of Cloud resources and, even if we consider data storage as virtually separated elements, security issues exist if privilege escalation allows for gaining control on any data on physical hosts. In this paper we show how it is possible to cope Model Driven Engineering techniques to security analysis and monitoring of Cloud infrastructures. For reducing overhead, we provide a formal profile of hosts thermal behaviors. Depending on services input workloads, we detect and forecast malicious actions by comparisons with real thermal data.

Task Deadline-Aware Energy-Efficient Scheduling Model for a Virtualized Cloud

Data centers in cloud environment consume high amount of energy which not only raises the electricity bills of the data center hosting organizations but also has the strong environmental footprints. Therefore, energy efficiency of the data centers has become an important research issue. Many energy efficiency approaches have been proposed in the literature for cloud. Efficient resource scheduling is one of the important approaches to achieve energy efficiency in cloud. In this paper, a task deadline-aware energy-efficient scheduling model for virtualized cloud is presented. Independent and dynamically arriving deadline-aware tasks are scheduled by virtualizing the physical hosts in the data center. The proposed scheduling model at the first instance achieves the energy efficiency by executing maximum workload in the operational state of the host and at the second instance by maximum energy saving in the idle state of the host. In the operational state of the host, maximum workload is executed by exploiting the task slack time in a new context, and in the idle state of the host, maximum energy is saved by deploying core-level granularity of dynamic voltage and frequency scaling. The presented scheduling model is evaluated on the synthetic and real-world workload. Results clearly indicate that the presented scheduling model outperforms the existing scheduling model on the account of performance parameters of guarantee ratio, total energy consumption, energy consumption per task and resource utilization.

Optimal multi-dimensional vector bin packing using simulated evolution

The use of the evolutionary heuristic simulated evolution for the optimization of the multi-dimensional vector bin packing problem, which is encountered in several industrial applications, is described. These applications range from production planning and steel fabrication to assignment of virtual machines (VMs) onto physical hosts at cloud-based data centers. The dimensions of VMs can include demands of CPU, memory, bandwidth, disk space etc. The generalized goodness functions that aid traversing the search space in an intelligent manner are designed to cater to the multidimensional nature of items (VMs). The efficiency of heuristics is tested by considering phase transition in the generation of difficult test cases. The quality of the heuristics is judged by determining how close the solution is to the estimated lower bound. A new implementation of a tighter lower bound is proposed. Experiments show that superior quality results are obtained by employing the proposed strategy.

Virtual machine placement for minimizing connection cost in data center networks

We consider an optimization problem for finding efficient placement of virtual machines (VMs) in a data center network. In this problem, we receive requests of VMs from customers, and seek to determine those physical machines in the network that host the requested VMs under capacity constraints. The objective of the problem is to minimize the total connection cost of the VM placement. We propose two models of the connection cost, called the centralized and the distributed models. In the former model, the connection cost for each request is defined as the minimum length of networks connecting all physical host machines and a specified root node, while the network does not connect the root node in the latter model. We present approximation algorithms for this optimization problem. For the centralized model, we present an O(logθ)-approximation algorithm, where θ is the ratio of the largest to the smallest requests. For the distributed model with uniform requests, we present an O(logn)-approximation algorithm for networks on n nodes. We also present a heuristic-based algorithm for the distributed model with non-uniform requests, and verify the performance of our algorithms through computational experiments.

Cloud Computing: Types, Topologies, Virtual Machine and VM Migration for decrease in power consumption

As the usage of internet is increasing day by day, more and more people are now getting connected with the web. In the last few years, internet has become a powerful platform that has changed the way to do a business and the way of communication. The number of users which were using the services of internet in last ten years is becoming double in the number, in every two years. As the number of users is increasing the more storage and new technologies and new platform to develop software are on a huge demand. Then cloud computing came into the picture, which is providing huge storage capacity, on demand self-service, pay according to usage, flexible pricing, resource pooling, software for users and services on their demand which can be accessed anywhere (SaaS), platform to develop new software (PaaS), any hardware which is required for development (IaaS). Now even XaaS i.e. everything as a service (here X is considered as everything, its value can be any) is also coming into the picture. To run one or more than one services or applications together but in their separated space or partition, we need virtual machine. It is refer to the one instance of an operating system (Guest OS). Then Virtualization makes it possible to run many operating systems and applications on the same physical computer at the same time. It makes the physical system energy efficient, as many machines or applications are able to run all together, in a way it consumes less energy. To make the physical system more energy efficient, migration of virtual machine is done, so that we can reduce the energy consumption. VM Migration is used for migration of virtual machine from one PM (physical machine) to another physical machine or host. In this paper, we will discuss about the introduction of cloud computing-its types and topologies; and the main focus would be on energy efficiency using VM migration.

Identifying Communication Patterns between Virtual Machines in Software-Defined Data Centers

Modern cloud data centers typically exploit management strategies to reduce the overall energy consumption. While most of the solutions focus on the energy consumption due to computational elements, the advent of the Software-Defined Network paradigm opens the possibility for more complex strategies taking into account the network traffic exchange within the data center. However, a network-aware Virtual Machine (VM) allocation requires the knowledge of data communication patterns, so that VMs exchanging significant amount of data can be placed on the same physical host or on low cost communication paths. In Infrastructure as a Service data centers, the information about VMs traffic exchange is not easily available unless a specialized monitoring function is deployed over the data center infrastructure. The main contribution of this paper is a methodology to infer VMs communication patterns starting from input/output network traffic time series of each VM and without relaying on a special purpose monitoring. Our reference scenario is a software-defined data center hosting a multi-tier application deployed using horizontal replication. The proposed methodology has two main goals to support a network-aware VMs allocation: first, to identify couples of intensively communicating VMs through correlation-based analysis of the time series; second, to identify VMs belonging to the same vertical stack of a multi-tier application. We evaluate the methodology by comparing different correlation indexes, clustering algorithms and time granularities to monitor the network traffic. The experimental results demonstrate the capability of the proposed approach to identify interacting VMs, even in a challenging scenario where the traffic patterns are similar in every VM belonging to the same application tier.

Layered virtual machine migration algorithm for network resource balancing in cloud computing

Due to the increasing sizes of cloud data centers, the number of virtual machines (VMs) and applications rises quickly. The rapid growth of large scale Internet services results in unbalanced load of network resource. The bandwidth utilization rate of some physical hosts is too high, and this causes network congestion. This paper presents a layered VM migration algorithm (LVMM). At first, the algorithm will divide the cloud data center into several regions according to the bandwidth utilization rate of the hosts. Then we balance the load of network resource of each region by VM migrations, and ultimately achieve the load balance of network resource in the cloud data center. Through simulation experiments in different environments, it is proved that the LVMMalgorithm can effectively balance the load of network resource in cloud computing.

CloudScout: A Non-Intrusive Approach to Service Dependency Discovery

Nowadays, numerous enterprises are migrating their applications into cloud computing environments. Typically, the applications are composed of several dependent service components that span many hosts and network devices. In light of this, exploring the dependency between service components can be beneficial for achieving fast network application response time. Moreover, it is significant to consolidate service components according to resource constraints, service dependency, and network structure. However, it is a tedious task to discover the dependency among service components without expert knowledge of the running application. In this paper, we propose CloudScout , a non-intrusive approach that is capable of automatically discovering dependent service components. CloudScout analyzes the correlation among service components based on the time-series information from system monitoring logs. We address two key challenges in CloudScout : service distance calculation and dependent service clustering. We conduct experiments on five applications with 290 service components that span 20 physical hosts across two data centers. The experimental results demonstrate that CloudScout can successfully discover the dependency among service components and facilitate reducing the network latency of network applications and distributed applications.

A Centralized Monitoring Mechanism in Virtualized Environments

Virtualization technology seals the compute and storage resources in the virtual machine to improve resource efficiency, which becomes the core support of cloud computing platform. Meanwhile, monitoring of virtual machines is also increasingly becoming an important security demand of the virtualization platform. Currently, the monitoring system of virtual machine behavior is usually deployed in the management domain on the same physical equipment of monitored virtual machine. From the perspective of system architecture, the distributed monitoring method is not conducive to a cloud platform provider for unified management of computing resources, and the monitoring behavior in the physical host of monitored virtual machine will consume a lot of limited computing and storage resources on the virtual machine server, thus occupying the limited service resources of the virtual machine. Since the observed monitoring overhead of a virtual machine lies mainly in information processing, a kind of centralized VMI mechanism (CVMI) is capable of dividing the monitoring mechanism, which acquires the status information of the virtual machine on the physical host of monitored virtual machine, and then sends information to the centralized security monitoring server, and analyzes the status information, in order to detect that the security events are operated on the centralized security monitoring server clusters. The experimental results show that, CVMI can effectively monitor the behavior of the virtual machine, and reduce the performance loss caused by introduction of performing monitoring in deployment of physical nodes of virtual machine.