Abstract:Driven by technology advances and economic forces, data centers are being built around the world to serve as the infrastructures for big data analysis services (e.g., GFS, Map-reduce, and Dryad), Internet-based applications (e.g., web search, e-commerce, and online social networking) and cloud computing services (e.g., Amazon EC2, Microsoft Windows Azure, and Google App Engine). Typically, a data center may contain thousands, tens of thousands or even millions of servers with significant inter-node bandwidth requirement between them. How to design efficient and sustainable data center network architectures, and how to manage/configure the networks are two fundamental challenges for data center networking (DCN) research, which form the basis of this Ph.D. dissertation work. Our research focuses on designing efficient DCN from both the architecture perspective and the management perspective. This dissertation is mainly composed of two key components called OSA and DAC. Specifically, OSA is a novel optical switching architecture for data center networks. At its core, by exploiting runtime reconfigurable optical devices, OSA flexibly adapts its topology and link capacities to various traffic patterns, thus achieving high network bisection bandwidth. The key idea we hope to deliver via OSA is that: a thin but flexible architecture, instead of static fat ones, can better support data center networking in a simple, cost-effective and sustainable way. On the other hand, DAC is a management system that can perform address autoconfiguration for both OSA network and other generic data center networks. DAC is motivated by the key observation that data center networks encode locality and topology information into their server and switch addresses for performance, reliability and routing purposes, making the traditional address autoconfiguration protocols, like DHCP, not applicable in the new data center environment. The core of DAC is to reduce the data center address autoconfiguration problem to the graph isomorphism (GI) problem and solve it efficiently by leveraging the data center network topology characteristics. Our designs have been evaluated and validated via extensive analysis, simulations and implementations. Particularly, we have prototyped OSA using the real optical devices and implemented DAC as an application over the Windows network stack.

SADC:Smart Architecture Data Center

A Scalable, Adaptive, Self-management and Fault-Tolerant Architecture for Digital Library

DDC: A Vision for a Disaggregated Datacenter

Data Center Sustainability: Revisits and Outlooks

A Brief Introduction to Infrastructure Planning for Next-Generation Smart Computing Data Centers

Virtual Machine Based Energy-Efficient Data Center Architecture for Cloud Computing: A Performance Perspective.

Future Data Center Energy-Conservation and Emission-Reduction Technologies in the Context of Smart and Low-Carbon City Construction

Smart-DRS: A Strategy of Dynamic Resource Scheduling in Cloud Data Center

GreenDCN: a General Framework for Achieving Energy Efficiency in Data Center Networks

Data center integrated energy system for sustainability: Generalization, approaches, methods, techniques, and future perspectives

Flexible, highly scalable and cost-effective network structures for data centers

Generic and Automatic Address Configuration for Data Center Networks

Energy-saving model for SDN data centers

Architecture design and management for data center networks

Integration Of Data Center Into The Distributed Satellite Cluster Networks: Challenges, Techniques, And Trends

Energy saving research on air conditioning system in data centers

Research on New Electric Power Data Center Based on DC Power Supply and Distribution Technology

A Low-Complexity and Energy-Efficient Dynamical Adjustment Strategy for Servers in a Data Center

An AI Model Automatic Training and Deployment Platform Based on Cloud Edge Architecture for DC Energy-Saving

Server Consolidation: An Approach to make Data Centers Energy Efficient and Green