Figure 1 - uploaded by Daniel de Souza Dias
Content may be subject to copyright.
Source publication
Data centers rely on virtualization to provide different services over a shared infrastructure. The placement of the different services and tasks in the physical machines is crucial for the performance of the whole system. A misplaced service can overload some network links, lead to congestion, or even connection disruptions. On the other hand, vir...
Contexts in source publication
Context 1
... the core of the network is under-provisioned while the bandwidth of top-of-rack switches may be wasted. Figure 1 illustrates a three-tier topology with 2 core switches, 4 aggregation switches, and 8 ToR switches. ...
Context 2
... virtual machine or a group of them, can only be instantiated in a server with enough free CPU cores and memory. Figure 1 illustrates our concept of partition in a three-tier topology. The dashed line over the servers and the top of rack switch form a partition of the data center with high degree of communication, since all the servers are connected to the same switch, being able to exchange data at full link speeds. ...
Similar publications
Recommending an optimized path for a large crowd poses a unique challenge to existing routing algorithms due to the interactions between users and the dynamic changes over road networks. Industries, researchers and end users show an enormous interest in crowdsourced data comprising social networks and user-generated content to remain updated with t...
![Fig. 1 LTE HeNB Standard Architecture [3]](profile/Aawatif-Hayar/publication/308066850/figure/fig1/AS:406061212291072@1473823945079/LTE-HeNB-Standard-Architecture-3_Q320.jpg)
Femtocells are small base stations designed for use as an ideal solution to ensure good radio coverage in the residential and corporate environments. In this paper we review LTE femtocell, LTE HeNB standard architecture and we detail our proposed architecture. It focuses more particularly on the options offered by the HeNB Energy Congestion Server...
This paper addresses several aspects of network level Quality of Service (QoS) provisioning for the most significant theoretical traffic scheduling algorithm, Generalized Processor Sharing (GPS). Various mechanisms supporting QoS in the deterministic regime such as per-session traffic characterization and delay bound computation as well as admissio...
Internet of vehicles (IoV), or vehicles to everything (V2X), is a relatively new concept for all vehicles connected to the internet. In this paper, we propose a solution to the problems of traffic congestion. Road traffic congestion has continued to be a major problem in many developed countries. This concept shows the place or location traffic con...
Because of the limited resources on radio access channels of third generation partnership projection (3GPP) network, one of the most challenging tasks posted by 3GPP cellular-based machine type communications (MTC) is congestion due to massive requests for connection to radio access network (RAN). In this paper, an overload control algorithm in 3GP...
Citations
... Researchers have suggested a dynamic VM placement algorithm in [23], [24] ...
The demand for computing resources like processing speed, memory, storage, and bandwidth is continually rising in cloud data centres. Cloud service providers have put up data centres across the world to meet the resource requirement of web applications. Thousands of extremely powerful servers are often installed in these data centres. The server virtualization approach is used to allocate resources during the deployment and provisioning of applications. In the world of cloud computing, server virtualization has emerged as a crucial component. It makes it possible to run multiple Virtual Machines (VMs) concurrently on top of a single Physical Machine (PM) or Server. Researchers have developed different VM placement and consolidation algorithms with trade-off between various conflicting performance parameters such as power consumption, SLA (Service Level Agreement) violation, application response time, number of VM migrations etc. Several researchers have researched this issue and offered a number of strategies to improve the values of one or more competing performance parameters. This research paper presents taxonomy and extensive systematic survey of several research papers on virtualized resource management in cloud environment to meet these performance parameters.
... Several papers study the problem of minimizing network traffic [6], [18], [8] to enhance the performance of a DC by selecting the most suitable physical machines for virtual machines. Daniel et al. [9] present a VMP algorithm to reallocate virtual machines in DC Server contingent on the memory usage, the traffic matrix network, and the overall CPU. The first phase of this VMP algorithm considers collecting data from VMs and DC topology. ...
Cloud computing is an innovative process that delivers on-demand services over the internet. Virtualization is considered as the key concept of cloud computing since it handles running multiple virtual resources in a single physical resource. Mapping the virtual machine (VM) to the appropriate physical machine (PM) is called virtual machine placement (VMP). In this context, the dilemma of placing VMs in the cloud environment presents a significant challenge that has been wholly addressed but not yet totally fixed. This paper provides a multi-objective decision-making approach for VMP in a cloud computing infrastructure. We propose a conic scalarization method to solve the optimization problem. Simulation results prove that the offline algorithm yields good results compared to online deterministic algorithms.
... To achieve this objective, they propose a two-tier heuristic algorithm which places VMs with large mutual traffic on PMs with low-cost connections. The authors in [32] propose a graph community-based algorithm for mapping VMs to PMs in DCNs. The algorithm first makes partitions of PMs regrading to their connectivity, then clusters VMs based on the amount of traffic exchanged between them. ...
Recent telecommunication paradigms, such as big data, Internet of Things (IoT), ubiquitous edge computing (UEC), and machine learning, are encountering with a tremendous number of complex applications that require different priorities and resource demands. These applications usually consist of a set of virtual machines (VMs) with some predefined traffic load between them. The efficiency of a cloud data center (CDC) as a prominent component in UEC significantly depends on the efficiency of its VM placement algorithm applied. However, VM placement is an NP-hard problem and thus there exist practically no optimal solution for this problem. In this paper, motivated by this, we propose a priority, power and traffic-aware approach for efficiently solving the VM placement problem in a CDC. Our approach aims to jointly minimize power consumption, network consumption and resource wastage in a multi-dimensional and heterogeneous CDC. To evaluate the performance of the proposed method, we compared it to the state-of-the-art on a fat-tree topology under various experiments. Results demonstrate that the proposed method is capable of reducing the total network consumption up to 29%, the consumption of power up to 18%, and the wastage of resources up to 68%, compared to the second-best results.
... According to the literature [4,6,12,13,23], there are a number of approaches proposed to solve this problem. ...
... The proposed model achieved its targets by improving the developed cloud application availability and minimizing its running cost. Dias et al. [6] proposed a bin-packing scheduling algorithm to reduce traffic communication and traffic cost. The proposed algorithm targets to group the highly communicated VMs into clusters then mapping these clusters to several servers. ...
Cloud computing has been considered a core model of elastic on-demand resource allocation using a pay-as-you go model. One of the big challenges of this environment is to provide high quality service (QoS) through efficient and stringent management of cloud data center resources. With the increasing demand for cloud based services, the traffic volume inside cloud data centers (DC) has been increased exponentially. Accordingly, and to provide high QoS, a proper scheduling mechanism has to be followed by the cloud service provider. Furthermore, accurate scheduling is necessary for advancing the problem of energy consumption and resource utilization. In this paper, we propose an optimal resource allocation and consolidation virtual machine (VM) placement model for multi-tier applications in modern large cloud DCs. The proposed model targets to optimize the DCs’ energy and communication cost that influence the overall cloud performance through Software Defined Networking (SDN) control features. To solve the formulated multi-objective optimization problem, a novel adaptive genetic algorithm is proposed. The experimental results validate the efficacy of the proposed model through extensive simulations using synthetic and real workload traces. These results show that the proposed model jointly optimizes cloud QoS as well as energy consumption.
... To achieve this objective, they propose a two-tier heuristic algorithm which places VMs with large mutual traffic on PMs with low-cost connections. The authors in [32] propose a graph community-based algorithm for mapping VMs to PMs in DCNs. The algorithm first makes partitions of PMs regrading to their connectivity, then clusters VMs based on the amount of traffic exchanged between them. ...
Recent telecommunication paradigms, such as big data, Internet of Things (IoT), ubiquitous edge computing (UEC), and machine learning, are encountering with a tremendous number of complex applications that require different priorities and resource demands. These applications usually consist of a set of virtual machines (VMs) with some predefined traffic load between them. The efficiency of a cloud data center (CDC) as a prominent component in UEC significantly depends on the efficiency of its VM placement algorithm applied. However, VM placement is an NP-hard problem and thus there exist practically no optimal solution for this problem. In this paper, motivated by this, we propose a priority, power and traffic-aware approach for efficiently solving the VM placement problem in a CDC. Our approach aims to jointly minimize power consumption, network consumption and resource wastage in a multi-dimensional and heterogeneous CDC. To evaluate the performance of the proposed method, we compared it to the state-of-the-art on a fat-tree topology under various experiments. Results demonstrate that the proposed method is capable of reducing the total network consumption up to 29%, the consumption of power up to 18%, and the wastage of resources up to 68%, compared to the second-best results.
... However, this model doesn't address communication dependency as well as DCN energy consumption. In (Dias & Costa, 2012), Dias et al. proposed a bin-packing scheduling algorithm to reduce the traffic cost. His method based on grouping the highly communicated VMs into a number of clusters and then mapping these clusters to partitions of physical machines. ...
the increasing demand for cloud computing services has led to the adoption of large-scale cloud data centers (DCs) to meet the user’s requirements. Efficiency and managing of such DCs have become a challenging problem. Consequently, energy-efficient solutions to optimize the whole DC energy consumption, optimize the application’s performance and reduce the cloud provider operational cost are crucial and needed. This paper addressed the problem of Virtual Machines (VMs) placement of multi-tier applications to maximize the compute resources utilization, minimize energy consumption, and reduce network traffic inside modern large-scale cloud DCs. The VM placement problem with communication dependencies among the VMs is modeled as an optimization problem. In this context, to solve the proposed problem, that formulated as a variant of a multiple knapsack problem, an adaptive genetic algorithm is implemented to find a near-optimal solution to the NP-complete modeled optimization problem. To validate the efficacy of the proposed model, extensive simulations are conducted using CloudSimSDN simulator. The experimental results validate the usefulness of the proposed model and its effectiveness in reducing DC energy consumption and optimize network traffic inside DC.
... Thus, the amount of data communicated between each pair of VMs is synthetically generated using a non-uniform distribution. Basically, data communication between a pair of VMs is modeled by a log-normal distribution [102] as 80% of the VMs have 800 kB/min traffic among each other while 4% of the VMs have 10 times higher traffic [103]. ...
... Data communication between a pair of VMs is modeled by a log-normal distribution [102]. ...
... Arrival and life-time of each VM, given in time slots, are generated by poisson and exponential distributions, respectively. Data correlation between each pair of VMs is generated by a lognormal distribution with the mean of 10 MB and uniform variance selection in the range of [1,4] [102]. For migration, the size of the VMs are in the range of 2, 4, and 8 GB according to the distribution of 60%, 30%, and 10%. ...
... Most studies on online scheduling only focus on the communication. [9] proposes an online virtual machine placement scheme based on re-allocation to improve the traffic distribution. This paper uses online migration to reduce the traffice congestion during the communication of the virtual machines, but online migration would produce a high cost and influence the placement of other users. ...
... We utilize a multi-rooted tree topology to represent the data center network in which each layer has the same capacities of the physical machines and physical links [7][8][9]. Let s G denote the data center network, such that ...
The rapid development of cloud computing and high requirements of operators requires strong support from the underlying Data Center Networks. Therefore, the effectiveness of using resources in the data center networks becomes a point of concern for operators and material for research. In this paper, we discuss the online virtual-cluster provision problem for multiple tenants with an aim to decide when and where the virtual cluster should be placed in a data center network. Our objective is maximizing the total revenue for the data center networks under the constraints. In order to solve this problem, this paper divides it into two parts: online multi-tenancy scheduling and virtual cluster placement. The first part aims to determine the scheduling orders for the multiple tenants, and the second part aims to determine the locations of virtual machines. We first approach the problem by using the variational inequality model and discuss the existence of the optimal solution. After that, we prove that provisioning virtual clusters for a multi-tenant data center network that maximizes revenue is NP-hard. Due to the complexity of this problem, an efficient heuristic algorithm OMS (Online Multi-tenancy Scheduling) is proposed to solve the online multi-tenancy scheduling problem. We further explore the virtual cluster placement problem based on the OMS and propose a novel algorithm during the virtual machine placement. We evaluate our algorithms through a series of simulations, and the simulations results demonstrate that OMS can significantly increase the efficiency and total revenue for the data centers.
... In [36], Dias et al. propose an online VMP algorithm to allocate and relocate VMs based on the analysis of usage pat- Relying on graph theory, the correlated VMs are aggregated and allocated to servers chosen based on the distance to each other such that the traffic congestion is reduced. With the goal of achieving minimum traffic congestion, a solution is proposed as the combination of a modified Girvan-Newman algorithm and allocation scheme specifications. ...
Cloud computing and network slicing are essential concepts of forthcoming 5G mobile systems. Network slices are essentially chunks of virtual computing and connectivity resources, configured and provisioned for particular services according to their characteristics and requirements. The success of cloud computing and network slicing hinges on the efficient allocation of virtual resources (e.g. VCPU, VMDISK) and the optimal placement of Virtualized Network Functions (VNFs) composing the network slices. In this context, this paper elaborates issues that may disrupt the placement of VNFs and VMs. The paper classifies the existing solutions for VM Placement (VMP) based on their nature, whether the placement is dynamic or static, their objectives, and their metrics. The paper then proposes a classification of VNF Placement (VNFP) approaches, first, regarding the general placement and management issues of VNFs, and second, based on the target VNF type.
... We tackle the semi-online formulation of the on-demand workload consolidation where tasks have to be allocated to servers in real-time. While the vast majority of the work carried on workload consolidation considers either the offline [12,13] or online [14,15,16] setting, we cast the problem in a semi-online framework by considering a short period of a few seconds within which tasks are grouped before being allocated to hosts. More precisely, we build upon the offline workload consolidation problem discussed in [17]. ...
Satisfying on-demand access to cloud computing infrastructures under quality-of-service constraints while minimising the wastage of resources is an important challenge in data centre resource management. In this paper we tackle this challenge in a semi-online workload management system allocating tasks with uncertain duration to physical servers. Our semi-online framework, based on a bin packing approach, allows us to gather information on incoming tasks during a short time window before deciding on their assignments. Our contributions are as follows: (i) we propose a formal framework capturing the semi-online consolidation problem; (ii) we propose a new dynamic and real-time allocation algorithm based on the incremental merging of bins; and (iii) an adaptation of standard bin packing heuristics with a local search algorithm for the semi-online context considered here. We provide a systematic study of the impact of varying time-period size and varying the degrees of uncertainty on the duration of incoming tasks. The policies are compared in terms of solution quality and solving time on a data-set extracted from a real-world cluster trace.
Our results show that, around periods of high demand, our best policy saves up to 40% of the resources compared to the other polices, and is robust to uncertainty in the task durations. Finally, we show that small increases in the allowable time window allows a significant improvement, but that larger time windows do not necessarily improve resource usage for real world datasets.
