Quantitative comparisons of the state‐of‐the‐art data center architectures

Bilal, Kashif; Khan, Samee U.; Zhang, Limin; Hayat, Khizar; Madani, Sajjad A.; Min‐Allah, Nasro; Wang, Lizhe; Chen, Dan; Iqbal, Majid; Xu, Cheng‐Zhong; Zomaya, Albert Y.

doi:10.1002/cpe.2963

Cited by 88 publications

(50 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The legacy DCN infrastructure inherently lacks the capability to meet the current growth trend and bandwidth demands. The legacy DCN architectures mainly suffer from: (a) energy-inefficiency, (b) poor scalability, (c) high cost, (d) low cross-section bandwidth, and (e) non-agility [6]. Many new DCN architectures, such as the fat-tree, DCell, VL2, BCube, flattened butterfly, and FiConn are proposed in response to the limitations posed by the legacy DCN architectures [6].…”

Section: Dcn Architecturesmentioning

confidence: 99%

“…The legacy DCN architectures mainly suffer from: (a) energy-inefficiency, (b) poor scalability, (c) high cost, (d) low cross-section bandwidth, and (e) non-agility [6]. Many new DCN architectures, such as the fat-tree, DCell, VL2, BCube, flattened butterfly, and FiConn are proposed in response to the limitations posed by the legacy DCN architectures [6]. Moreover, various hybrid DCN architectures using amalgamation of optical and wireless technologies have been proposed in recent years [14].…”

Section: Dcn Architecturesmentioning

confidence: 99%

“…The state of the art DCN architectures mainly focus on: (a) performance, (b) reliability, (c) fault tolerance, (d) high end-to-end bandwidth, and (e) agility. As a consequence, the new DCN architectures are often highly overprovisioned and underutilized [6], which triggers the inquest to design newer energy-efficient architectures and techniques.…”

Section: Dcn Architecturesmentioning

confidence: 99%

“…The server-centric models place the routing and switching capability within the computational servers. Computational servers are used for packet forwarding and routing besides performing computational tasks [6].…”

Section: Dcn Architecturesmentioning

confidence: 99%

“…To deliver 1:1 oversubscription ratio, the commodity network switches are arranged in a Clos topology within the fat-tree. The fat-tree architecture delivers high end-to-end bandwidth at a much lower cost and energy consumption [6]. Subsequently proposed successors of the fat-tree, such as VL2 and Portland, also use the Clos structure and commodity network equipment [7].…”

Section: Dcn Architecturesmentioning

confidence: 99%

See 4 more Smart Citations

Green Data Center Networks: Challenges and Opportunities

Bilal

Khan

Zomaya

2013

2013 11th International Conference on Frontiers of Information Technology

Self Cite

View full text Add to dashboard Cite

Abstract-The concerns about environmental impacts, electricity cost, and energy needs of data centers are mounting. These concerns have given significant impetus to consider energy efficiency as one of the major data center design parameters for researchers and IT executives. Tremendous growth in the size and computing demands of data centers mandate scalable, fault tolerant, and energy efficient network infrastructure within data centers. In the recent years, new Data Center Network (DCN) architectures have aptly proposed a viable solution to the issues posed by the legacy DCNs. New DCN architectures are required to be energy efficient and energy proportional, besides dealing with other identified design drawbacks, such as scalability, fault tolerance, end-to-end bandwidth, and non-agility. In this article, we will deliberate on the various DCN architectures, their inherent problems, and the potentials to achieve energy efficiency. We will also shed light on myriad techniques that may be suitable to make the DCN architectures more energy efficient and energy proportional. Finally, we will outline future directions.

show abstract

Section: Dcn Architecturesmentioning

confidence: 99%

Section: Dcn Architecturesmentioning

confidence: 99%

Section: Dcn Architecturesmentioning

confidence: 99%

Section: Dcn Architecturesmentioning

confidence: 99%

Section: Dcn Architecturesmentioning

confidence: 99%

See 3 more Smart Citations

Green Data Center Networks: Challenges and Opportunities

Bilal

Khan

Zomaya

2013

2013 11th International Conference on Frontiers of Information Technology

Self Cite

View full text Add to dashboard Cite

show abstract

Scalable optimization in grid, cloud, and intelligent network computing – foreword

Kołodziej

Khan

Talbi

2013

Concurrency and Computation

View full text Add to dashboard Cite

Global optimization in large-scale distributed systems requires massive amounts of computations for complex objective functions. Conventional global optimization based on stochastic algorithms cannot guarantee an actual global optimum with a finite searching iteration. Therefore, scalability is a desirable feature for the optimization techniques in highly distributed dynamic environments, where the storage and computing capabilities can be spread over a wide geographical area. They must dynamically adapt to organizational relationships and real-world uncertainties.Intelligent Networks, such as grids, peer-to-peer, ad hoc networks, constellations, and clouds enable the flexible routing and charging, advanced user interactions and the aggregation and sharing of geographically distributed resources. Collectively owned and managed by distinct organizational bodies, such complex large-scale distributed systems typically encompass computational resources from different institutions, enterprises, and individuals and are governed by heterogeneous administrative policies and regulations. System management techniques must therefore be able to group, predict, and classify different sets of rules, configuration directives, and environmental conditions to impose dissimilar usage policies on various users and resources. They must effectively deal with various optimization criteria, users' requirements, massive data processing, and, finally, uncertainties in system information that may be incomplete, imprecise, and fragmentary. Next information technology architectures, such as green cloud-to-cloud systems and green mobile clouds, provide elastic and in fact unlimited resources, including storage, as various services to cloud users with possible minimal energy utilization. However, both cloud users and cloud service providers are almost certain to be from different trust domains. Therefore, a secure user-enforced data access control mechanism must be provided before cloud users have the liberty to outsource sensitive data to the cloud for storage and further processing.With the advent of intelligent networks, where efficient interdomain operation and high scalability of the whole system are the most important features, it is arguably required to investigate novel methods and techniques to enable secure access to data and resources, flexible communication, efficient scheduling, self-adaptation, decentralization, and self-organization.This special issue herewith presents six research papers with novel concepts in the analysis, implementation, and evaluation of the next generation of intelligent scalable techniques for data-intensive processing and global optimization problems in large-scale distributed systems.The first three papers discuss novel scalable solutions of data-intensive global optimization problems in well-known large-scale network environments. The presented techniques and their implementations are based on formal mathematical and logical models with the new optimization criteria (energy conservation), semantic rul...

show abstract

Doughnutie: An efficient and low‐latency cloud data center network architecture

Nasirian

Faghani

Farzanegan

2021

Concurrency and Computation

View full text Add to dashboard Cite

Data center network provides the supporting infrastructure for cloud computing. Torus interconnection network is widely employed in the cloud data center networks. However, its long network diameter and average path length (APL) are prohibitive factors for supporting the delay-sensitive applications. With the aim of meeting latency, network capacity, fault-tolerance, scalability, and energy-efficiency requirements of cloud data center networks, we propose a new data center network architecture, named Doughnutie. Doughnutie is built by adding a number of flyway connections to the base Torus network to ameliorate its performance. Doughnutie does not follow a single principle for establishing all of its flyways and consequently, each part of the network can profit from all local, faraway, and neither-local-nor-faraway types of connection. Doughnutie decreases both network diameter and APL, which in turn leads to higher communication efficiency and lower network latency. Moreover, an exclusively designed loop-free load-balanced adaptive routing algorithm, named Doughnuter, is proposed.Employing the Doughnuter routing algorithm results in more efficient utilization of all the Doughnutie network resources. Theoretical analysis and simulation results also prove the superiorities of Doughnutie in throughput, load balancing, fault tolerance, and energy-efficiency.

show abstract

Quantitative comparisons of the state‐of‐the‐art data center architectures

Cited by 88 publications

References 33 publications

Green Data Center Networks: Challenges and Opportunities

Green Data Center Networks: Challenges and Opportunities

Scalable optimization in grid, cloud, and intelligent network computing – foreword

Doughnutie: An efficient and low‐latency cloud data center network architecture

Contact Info

Product

Resources

About