DPillar: Scalable Dual-Port Server Interconnection for Data Center Networks

Liao, Yong; Yin, Dong; Gao, Lixin

doi:10.1109/icccn.2010.5560132

Cited by 30 publications

(19 citation statements)

References 6 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Fig. 10(b) shows the same distributions of SWCube(9, 3) and SWKautz (12,3). In the figures, a bar represents the percentage of paths whose lengths are equal to the corresponding value.…”

Section: B Average Path Lengthmentioning

confidence: 52%

“…First, we notice that in existing works [1], [9]- [12], the lengths of a server-to-server-direct hop and a serverto-server-via-switch hop are assumed to be equal. We call this assumption the HOmogeneous Hop (HOH) assumption.…”

Section: Introductionmentioning

confidence: 99%

“…Thus, design and comparison based on this assumption may be inappropriate. For example, in FiConn [11] and BCN [1], there exist serverto-server-direct hops, while in DPillar [12], any two servers are not directly connected. In this paper, we propose the concept of Normalized Switch Delay (NSD), which is defined as the switch's packet forwarding delay divided by the server's forwarding delay (when they have no other load), to distinguish these two kinds of sever-to-server hops to unify the design of DCN architectures.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

On the design and analysis of Data Center Network architectures for interconnecting dual-port servers

2014

IEEE INFOCOM 2014 - IEEE Conference on Computer Communications

View full text Add to dashboard Cite

Abstract-We consider the design and analysis of Data Center Network (DCN) architectures for interconnecting dual-port servers. Unlike existing works, we propose the concept of Normalized Switch Delay (NSD) to distinguish a server-to-server-direct hop and a server-to-server-via-switch hop, to unify the design of DCN architectures. We then consider a fundamental problem: maximizing the number of dual-port servers, given network diameter and switch port number; and give an upper bound on this maximum number. Two novel architectures are proposed: SWCube and SWKautz, based on the generalized hypercube and Kautz graph, respectively, which in most cases accommodate more servers than BCN [1], which was claimed to be the largest known architecture. Compared with three existing architectures, SWCube and SWKautz demonstrate various advantages. Analysis and simulations also show that SWCube and SWKautz have nice properties for DCNs, such as low diameter, good fault-tolerance, and capability of efficiently handling network congestion.

show abstract

“…Fig. 10(b) shows the same distributions of SWCube(9, 3) and SWKautz (12,3). In the figures, a bar represents the percentage of paths whose lengths are equal to the corresponding value.…”

Section: B Average Path Lengthmentioning

confidence: 52%

Section: Introductionmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

On the design and analysis of Data Center Network architectures for interconnecting dual-port servers

2014

IEEE INFOCOM 2014 - IEEE Conference on Computer Communications

View full text Add to dashboard Cite

show abstract

“…A failure of the core or aggregation switches may degrade operation or even disable a large number of racks. Therefore, fat-tree architectures will be replaced with distributed approaches like DCell [24], BCube [25], FiConn [26], or DPillar [27] in future data centers.…”

Section: Data Center Architecturesmentioning

confidence: 99%

Simulation and Performance Analysis of Data Intensive and Workload Intensive Cloud Computing Data Centers

2012

View full text Add to dashboard Cite

Data centers are becoming increasingly popular for the provisioning of computing resources. The cost and operational expenses of data centers have skyrocketed with the increase in computing capacity [1]. Energy consumption is a growing concern for data center operators. It is becoming one of the main entries on a data center operational expenses (OPEX) bill [2,3]. The Gartner Group estimates energy consumptions to account for up to 10% of the current OPEX, and this estimate is projected to rise to 50% in the next few years [4]. However, computing-based energy consumption is not the only power-related portion of the OPEX bill. High power consumption generates heat and requires an accompanying cooling system that costs in a range of $2-$5 million per year for classical data centers [5]. Failure to keep data center temperatures within operational ranges drastically decreases hardware reliability and may potentially violate the service level agreement (SLA) with the customers.From the perspective of energy efficiency, a cloud computing data center can be defined as a pool of computing and communication resources organized in the way to transform the received power into computing or data transfer work to satisfy user demands. The first power saving solutions focused on making the data center hardware components power efficient. Technologies, such as dynamic voltage and frequency scaling (DVFS), and dynamic power management (DPM) [6], were

show abstract

“…Therefore, less number of switches is needed. Several proposals investigate the viability of servercentric DCNs by developing recursively defined DCN (e.g., DCell [24], BCube [25], FiConn [26], DPillar [27], and BCN [28]) or using a fixed topology DCN (e.g., CamCube [29] which uses torus topology). Most server-centric DCNs have improved scalability and cost-effectiveness as compared to most switchcentric DCNs [24]- [28].…”

mentioning

confidence: 99%

Wireless Communication in Data Centers: A Survey

Hamza

Deogun

Alexander

2016

IEEE Commun. Surv. Tutorials

103

View full text Add to dashboard Cite

Abstract-Data centers (DCs) is becoming increasingly an integral part of the computing infrastructures of most enterprises. Therefore, the concept of DC networks (DCNs) is receiving an increased attention in the network research community. Most DCNs deployed today can be classified as wired DCNs as copper and optical fiber cables are used for intra-and inter-rack connections in the network. Despite recent advances, wired DCNs face two inevitable problems; cabling complexity and hotspots.To address these problems, recent research works suggest the incorporation of wireless communication technology into DCNs. Wireless links can be used to either augment conventional wired DCNs, or to realize a pure wireless DCN. As the design spectrum of DCs broadens, so does the need for a clear classification to differentiate various design options. In this paper, we analyze the free space optical (FSO) communication and the 60 GHz radio frequency (RF), the two key candidate technologies for implementing wireless links in DCNs. We present a generic classification scheme that can be used to classify current and future DCNs based on the communication technology used in the network. The proposed classification is then used to review and summarize major research in this area. We also discuss open questions and future research directions in the area of wireless DCs.Index Terms-Wireless data centers, 60 GHz, free space optical (FSO), optical wireless communication (OWC), data centers, data center network.

show abstract

DPillar: Scalable Dual-Port Server Interconnection for Data Center Networks

Cited by 30 publications

References 6 publications

On the design and analysis of Data Center Network architectures for interconnecting dual-port servers

On the design and analysis of Data Center Network architectures for interconnecting dual-port servers

Simulation and Performance Analysis of Data Intensive and Workload Intensive Cloud Computing Data Centers

Wireless Communication in Data Centers: A Survey

Contact Info

Product

Resources

About