Homepage / Searching for: improving datacenter network performance via intelligent network edge

Welcome to Read Book Online

Download improving datacenter network performance via intelligent network edge or read improving datacenter network performance via intelligent network edge online books in PDF, EPUB and Mobi Format. Click Download or Read Online button to get improving datacenter network performance via intelligent network edge book now. Note:! If the content not Found, you must refresh this page manually.

Improving Datacenter Network Performance Via Intelligent Network Edge

DOWNLOAD
Author by : Keqiang He
Languange Used : en
Release Date : 2017
Publisher by :

ISBN : OCLC:1013189973

Datacenter networks are critical building blocks for modern cloud computing infrastructures. In this dissertation, we show how we can leverage the flexibility and high programmability of datacenter network edge (i.e., end-host networking) [101, 102] to improve the performance of three key functionalities in datacenter networks -- traffic load balancing, congestion control and rate limiting. Datacenter networks need to deal with a variety of workloads, ranging from latency-sensitive small flows to bandwidth-hungry large flows. In-network hardware-based load balancing schemes which are based on flow hashing, e.g., ECMP, cause congestion when hash collisions occur. To solve this problem, we propose a soft-edge load balancing scheme called Presto. Presto load-balances on near uniform-sized small data units (flowcells) and spreads flowcells across the symmetric network via the virtual switches on the senders. Because of fine-grained flowcell-level load balancing, packets may arrive out of order at the receiver side, so we propose a mechanism to handle reordering in the Generic Receive Offload (GRO) functionality below the TCP layer. Presto avoids the hash collision problem and improves traffic load balancing performance significantly. Optimized traffic load balancing alone is not sufficient to guarantee high-performance datacenter networks. Virtual Machine (VM) technology plays an integral role in modern multi-tenant clouds by enabling a diverse set of software to be run on a unified underlying framework. This flexibility, however, comes at the cost of dealing with outdated, inefficient, or misconfigured TCP stacks implemented in the VMs. We propose a congestion control virtualization technique called AC/DC TCP. AC/DC TCP exerts fine-grained control over arbitrary tenant TCP stacks by enforcing per-flow congestion control in the virtual switch (vSwitch) in the hypervisor. AC/DC TCP is light-weight, flexible, scalable and can police non-conforming flows. Besides queueing latency in network switches, we observe that rate limiters on end-hosts can also increase network latency by an order of magnitude or even more. To this end, we propose two techniques -- DEM and SPRING to improve the performance of rate limiters. Our experiment results demonstrate that DEM and SPRING-enabled rate limiters can achieve high stable throughput and low latency....

Improving Datacenter Performance With Network Offloading

DOWNLOAD
Author by : Yanfang Le
Languange Used : en
Release Date : 2020
Publisher by :

ISBN : OCLC:1245425211

There has been a recent emergence of distributed systems in datacenters, such as MapReduce and Spark for data analytics and TensorFlow and PyTorch for machine learning. These frameworks are not only computation and memory intensive, they also place high demands on the network for distributing data. The fast-growing Ethernet speed mitigates the high demand a bit. However, as Ethernet speed outgrows the CPU processing power, it not only requires us to rethink the existing algorithms for different network layers, but also provides opportunities to innovate with new application designs, such as datacenter resource disaggregation [3] and in-network computation applications [4, 5, 6]. The fast network devices come with a programmability feature, which enables offloading computation tasks from CPU to NICs or switches. Network offloading to programmable hardware is a promising approach to help relieve processing pressure on the CPU for computation-intensive applications, e.g., Spark, or reduce the network traffic for network-intensive applications, e.g., TensorFlow. However, leveraging programmable hardware effectively is challenging due to the limited memory capacity and restricted programming model. In order to understand how to leverage the advantage of network offloading in developing new network stacks, network protocols, and applications, the following question needs to be answered: how to do judicious division between the programmable hardware and software for network offload given limited resources and restricted programming models? Driven by the real application demand while exploring the answer to this question, we first propose RoGUE, a new congestion control and recovery mechanism for RDMA over Converged Ethernet that does not rely on PFC while preserving the benefits of running RDMA, i.e., low CPU and low latency. To preserve the low CPU benefit, RoGUE offloads packet pacing to the NIC. Though RoGUE achieves better performance in extensive testbed evaluations, the architecture for optimal congestion control should be a centralized packet scheduler [7], which has global visibility into packet reservation requests from all the servers. Given all the hosts are connected through switches and the emerging programmable switch hardware can have stateful objects, we designed a centralized packet scheduler at the switch, called PL2, to provide stable and near-zero-queuing in the network by proactively reserving switch buffers for packet bursts in the appropriate time-slots. Congestion control is an essential component in the networking stack because application demand for the network is higher than link speed. To eliminate the net- work congestion control, the fundamental solution is reducing the network traffic such that the application demand for the network is no more than link speed. We observed that we are able to reduce the network traffic for distributed training sys- tems by offloading a critical function, gradients aggregation, to the programmable switch. Each worker in the distributed training system sends gradients over the network to special components, parameter servers, to do aggregation, which is a simple add operator. Thus, we propose ATP, a network service for in-network aggregation aimed at modern multi-rack, multi-job DT settings. ATP performs decentralized, dynamic, best-effort aggregation, enables efficient and equitable sharing of limited switch resources across simultaneously running DT jobs, and gracefully accommodates heavy contention for switch resources....

Artificial Intelligence For Cloud And Edge Computing For Super Networks 5g How To Monetize 5g Super Networks For Cloud And Edge Computing Using Ai

DOWNLOAD
Author by : Sajjad Ahmad
Languange Used : en
Release Date : 2024-04-01
Publisher by : Sajjad ahmad

ISBN : 9798321323854

"Artificial Intelligence for Cloud and Edge Computing for Super Networks -5G" Harnessing 5G and Edge Cloud Computing for Business Innovation is a comprehensive guidebook that delves into the transformative potential of edge cloud computing in conjunction with 5G networks. Authored by industry experts, the book offers a detailed exploration of how these cutting-edge technologies intersect to revolutionize various business sectors. From healthcare and industrial automation to sports venues and entertainment, the book provides insightful use cases, real-life examples, and practical strategies for leveraging edge computing to drive innovation, enhance operational efficiency, and unlock new revenue streams. With a focus on business-to-business applications, the book serves as a roadmap for organizations seeking to capitalize on the power of edge computing and 5G to stay ahead in today's digital landscape....

Build A Smarter Data Center With Juniper Networks Qfabric

DOWNLOAD
Author by : Bill White
Languange Used : en
Release Date : 2012-03-07
Publisher by : IBM Redbooks

ISBN : 9780738451190

In this IBM® RedguideTM document, we highlight the key requirements for a smarter data center network and show how the data center fabric, a new switching architecture, provides the required performance, scalability, and management. We explore Juniper Networks' QFabric, a revolutionary DCN fabric product, and describe how its characteristics and key network innovations provide real business value in rapid service deployment, cost-efficient service delivery, energy efficiency, and business resiliency and security. We examine Juniper's QFabric design, product software, hardware, and deployment, and illustrate how QFabric can drastically improve your DCN while reducing your business costs. We describe three common QFabric network use cases that highlight fundamental changes in DCN architecture. Use cases are based on our project experiences, specifically optimized application delivery control, secure isolation provisioning of a multi-tenant environment, and support of business continuity. IBM understands that the first step in transforming network infrastructure is developing an enterprise network architecture that considers business and IT environments, security and privacy policies, service priorities, and growth plans. This guide describes how to migrate to a smarter data center using QFabric and also considers organizational aspects of migration. Over decades, IBM has built deep technical expertise and understanding of the evolving demands of network, server, storage, and desktop virtualization. IBM has extensive design and integration experience in complex DCN infrastructures and cloud computing environments. And IBM has a global pool of skilled networking professionals with in-depth IT and networking infrastructure knowledge and world class project management skills. IBM and Juniper Networks' strong partnership offers leading edge network products and technologies that will help you create and implement this unrivalled DCN design using the information covered in this paper....

Introduction To Storage Area Networks

DOWNLOAD
Author by : Jon Tate
Languange Used : en
Release Date : 2018-10-09
Publisher by : IBM Redbooks

ISBN : 9780738442884

The superabundance of data that is created by today's businesses is making storage a strategic investment priority for companies of all sizes. As storage takes precedence, the following major initiatives emerge: Flatten and converge your network: IBM® takes an open, standards-based approach to implement the latest advances in the flat, converged data center network designs of today. IBM Storage solutions enable clients to deploy a high-speed, low-latency Unified Fabric Architecture. Optimize and automate virtualization: Advanced virtualization awareness reduces the cost and complexity of deploying physical and virtual data center infrastructure. Simplify management: IBM data center networks are easy to deploy, maintain, scale, and virtualize, delivering the foundation of consolidated operations for dynamic infrastructure management. Storage is no longer an afterthought. Too much is at stake. Companies are searching for more ways to efficiently manage expanding volumes of data, and to make that data accessible throughout the enterprise. This demand is propelling the move of storage into the network. Also, the increasing complexity of managing large numbers of storage devices and vast amounts of data is driving greater business value into software and services. With current estimates of the amount of data to be managed and made available increasing at 60% each year, this outlook is where a storage area network (SAN) enters the arena. SANs are the leading storage infrastructure for the global economy of today. SANs offer simplified storage management, scalability, flexibility, and availability; and improved data access, movement, and backup. Welcome to the cognitive era. The smarter data center with the improved economics of IT can be achieved by connecting servers and storage with a high-speed and intelligent network fabric. A smarter data center that hosts IBM Storage solutions can provide an environment that is smarter, faster, greener, open, and easy to manage. This IBM® Redbooks® publication provides an introduction to SAN and Ethernet networking, and how these networks help to achieve a smarter data center. This book is intended for people who are not very familiar with IT, or who are just starting out in the IT world....

Edge Intelligence In The Making

DOWNLOAD
Author by : Sen Lin
Languange Used : en
Release Date : 2022-06-01
Publisher by : Springer Nature

ISBN : 9783031023804

With the explosive growth of mobile computing and Internet of Things (IoT) applications, as exemplified by AR/VR, smart city, and video/audio surveillance, billions of mobile and IoT devices are being connected to the Internet, generating zillions of bytes of data at the network edge. Driven by this trend, there is an urgent need to push the frontiers of artificial intelligence (AI) to the network edge to fully unleash the potential of IoT big data. Indeed, the marriage of edge computing and AI has resulted in innovative solutions, namely edge intelligence or edge AI. Nevertheless, research and practice on this emerging inter-disciplinary field is still in its infancy stage. To facilitate the dissemination of the recent advances in edge intelligence in both academia and industry, this book conducts a comprehensive and detailed survey of the recent research efforts and also showcases the authors' own research progress on edge intelligence. Specifically, the book first reviews the background and present motivation for AI running at the network edge. Next, it provides an overview of the overarching architectures, frameworks, and emerging key technologies for deep learning models toward training/inference at the network edge. To illustrate the research problems for edge intelligence, the book also showcases four of the authors' own research projects on edge intelligence, ranging from rigorous theoretical analysis to studies based on realistic implementation. Finally, it discusses the applications, marketplace, and future research opportunities of edge intelligence. This emerging interdisciplinary field offers many open problems and yet also tremendous opportunities, and this book only touches the tip of iceberg. Hopefully, this book will elicit escalating attention, stimulate fruitful discussions, and open new directions on edge intelligence....

Improving Fault Tolerance And Performance Of Data Center Networks

DOWNLOAD
Author by : Vincent Liu
Languange Used : en
Release Date : 2016
Publisher by :

ISBN : OCLC:981509536

Data center networks are a key component to the explosive growth of cloud computing---enabling the utilization of tens to hundreds of thousands of co-located servers for large-scale computing and services. As applications and data sets continue to grow rapidly, the challenge for data center networks is to keep pace---by providing enough bandwidth while also lowering costs, increasing flexibility, and maintaining reliability. My thesis is that a key part of the answer is the network's wiring topology: topology has foundational cross-layer effects, and a small amount of intentional asymmetry in the topology can help data center networks meet that challenge. I present two complementary innovations that demonstrate this. The first, F10, is a co-design of the network topology and failover protocols to provide efficient, near-instantaneous, fine-grained, and localized recovery and rebalancing for common-case network failures. My results show that following network link and switch failures, F10 has 1/7th the packet loss of current schemes. The second innovation, Subways, proposes and evaluates a new method to add network capacity by connecting multiple network links per server in an overlapping topology. Using a simulation-based methodology, my work shows that Subways offers substantial performance benefits for popular application workloads: up to a 3.1x speedup in MapReduce and a 2.5x throughput improvement in memcache for a fixed average request latency, relative to an equivalent-bandwidth network that differs only in its wiring....

Heterogenous Computational Intelligence In Internet Of Things

DOWNLOAD
Author by : Pawan Kumar Singh
Languange Used : en
Release Date : 2023-11
Publisher by :

ISBN : 103242639X

"We have seen a sharp increase in the development of data transfer techniques in the networking industry during the last few years. We can see that the photos are assisting clinicians in detecting Covid-19 infection in patients even in the current Covid-19 pandemic condition. With the aid of ML/AI, medical imaging, such as lung X-rays for Covid-19 infection, is crucial in the early detection of many diseases. We also learned that in the Covid-19 scenario, wired and wireless networking are improved for data transfer but have network congestion. An intriguing concept that has the ability to reduce spectrum congestion and continuously offer new network services is providing wireless network virtualization. The degree of virtualization and resource sharing varies between the paradigms. Each paradigm has both technical and non-technical issues that need to be handled before wireless virtualization becomes a common technology. For wireless network virtualization to be successful, these issues need careful design and evaluation. Future wireless network architecture must adhere to a number of Quality of Service requirements (QoS). Virtualization has been extended to wireless networks as well as conventional ones. By enabling multi-tenancy and tailored service with a wider range of carrier frequencies, it improves efficiency and utilization. In the IoT environment, wireless users are heterogeneous, and the network state is dynamic, making network control problems extremely difficult to solve as dimensionality and computational complexity keep rising quickly. Deep Reinforcement Learning (DRL) has been developed by the use of Deep Neural Networks (DNNs) as a potential approach to solve high-dimensional and continuous control issues effectively. Deep Reinforcement Learning techniques provide great potential in IoT, edge and SDN scenarios and are used in heterogeneous networks for IoT-based management on the QoS required by each Software Defined Network (SDN) service. While DRL has shown great potential to solve emerging problems in complex wireless network virtualization, there are still domain-specific challenges that require further study, including the design of adequate DNN architectures with 5G network optimization issues, resource discovery and allocation, developing intelligent mechanisms that allow the automated and dynamic management of the virtual communications established in the SDNs which is considered as research perspective"--...

Research Anthology On Developing And Optimizing 5g Networks And The Impact On Society

DOWNLOAD
Author by : Management Association, Information Resources
Languange Used : en
Release Date : 2020-11-27
Publisher by : IGI Global

ISBN : 9781799877547

As technology advances, the emergence of 5G has become an essential discussion moving forward as its applications and benefits are expected to enhance many areas of life. The introduction of 5G technology to society will improve communication speed, the efficiency of information transfer, and end-user experience to name only a few of many future improvements. These new opportunities offered by 5G networks will spread across industry, government, business, and personal user experiences leading to widespread innovation and technological advancement. What stands at the very core of 5G becoming an integral part of society is the very fact that it is expected to enrich society in a multifaceted way, enhancing connectivity and efficiency in just about every sector including healthcare, agriculture, business, and more. Therefore, it has been a critical topic of research to explore the implications of this technology, how it functions, what industries it will impact, and the challenges and solutions of its implementation into modern society. Research Anthology on Developing and Optimizing 5G Networks and the Impact on Society is a critical reference source that analyzes the use of 5G technology from the standpoint of its design and technological development to its applications in a multitude of industries. This overall view of the aspects of 5G networks creates a comprehensive book for all stages of the implementation of 5G, from early conception to application in various sectors. Topics highlighted include smart cities, wireless and mobile networks, radio access technology, internet of things, and more. This all-encompassing book is ideal for network experts, IT specialists, technologists, academicians, researchers, and students....

Secure Edge And Fog Computing Enabled Ai For Iot And Smart Cities

DOWNLOAD
Author by : Ahmed A. Abd El-Latif
Languange Used : en
Release Date :
Publisher by : Springer Nature

ISBN : 9783031510977

...