Data Center Networks Research Articles

Fine-grained load balancing with proactive prediction and adaptive rerouting in data center

Though the existing load balancing designs successfully make full use of available parallel paths and attain high bisection network bandwidth, they reroute flows regardless of their dissimilar performance requirements. But traffic in modern data center networks exhibits short bursts characteristic, which can easily lead to network congestion. The short flows suffer from the problems of large queuing delay and packet reordering, while the long flows fail to obtain high throughput due to low link utilization and packet reordering. In order to solve these inefficiency, we designed a fine-grained load balancing method (FLB), which uses an active monitoring mechanism to split traffic, and flexibly transfers flowlet to non-congested path, effectively reducing the negative impact of burst flow on network performance. Besides, to avoid packet reordering, FLB leverages the probe packets to estimate the end-to-end delay, thus excluding paths that potentially cause packet reordering. The test results of NS2 simulation show that FLB significantly reduces the average and tail flow completion time of flows by up to 59% and 56% compared to the state-of-the-art multi-path transmission scheme with less computational overhead, as well as increases the throughput of long flow.

Retracted: Fault-Tolerant Secure Routing Based on Trust Evaluation Model in Data Center Networks

Retracted: Fault-Tolerant Secure Routing Based on Trust Evaluation Model in Data Center Networks

Exploring the benefits of using co-packaged optics in data center and AI supercomputer networks: a simulation-based analysis [Invited

We investigate the advantages of using co-packaged optics in next-generation data center and AI supercomputer networks. The increased escape bandwidth offered by co-packaged optics provides multiple possibilities for building 50T switches and beyond, expanding the opportunities in both the data center and supercomputing domains. This provides network architects with the opportunity to expand their design space and develop simplified networks with enhanced network locality properties. Co-packaging at the switch and server points enables networks with double capacity while reducing the switch count by 64% compared to state-of-the-art systems. We evaluate these concepts through discrete-event simulations using all-to-all and all-reduce traffic patterns that simulate collective communications commonly found in network-bound applications. Initially, we investigate the all-to-all overhead involved in distributing the virtual machines of the applications across multiple leaf switches and compare it to the scenario in which all VMs are placed under a single switch. Subsequently, we evaluate the performance of an AI supercomputing cluster by simulating both patterns for different message sizes, while also varying the number of participating nodes. The results suggest that networks with improved locality properties become increasingly important as the network stack operates at higher speeds; for a stack latency of 1.25 µs, placing the applications under multiple switches can result in up to 68% higher completion times than placing them under a single switch. For AI supercomputers, significant improvements are observed in the mean server throughput, reaching more than 90% for configurations involving 256 nodes and message sizes of at least 128 KiB.

Maximum elastic scheduling of virtual machines in general graph cloud data center networks

ABSTRACT In this research, we pioneer a novel method to evaluate the maximum admissible load (MAL) for virtual machines (VMs) in physical machines (PMs) in data centre networks (DCNs), without restricting DCN topologies. This unique approach simplifies the issue into a single-source, multiple-sink maximum flow problem. It also resolves the maximum elastic scheduling problem by determining the optimal load for consistent growth without reassigning tasks. An effective strategy for these challenges is introduced and validated through extensive simulations.

Enhanced Coordination in the PV–HESS Microgrids Cluster: Introducing a New Distributed Event Consensus Algorithm

To ensure reliable power delivery to customers under potential disturbances, the coordination of a microgrid cluster (MGC) is essential. Various control strategies—centralized, decentralized, distributed, and hierarchical—have been explored in the literature to achieve this goal. The hierarchical control method, with three distinct levels, has proven effective in fostering coordination among microgrids (MGs) within the cluster. The third control level, utilizing a time-triggering consensus protocol, relies on a continuous and reliable communication network for data exchange among MGs, leading to resource-intensive operations and potential data congestion. Moreover, uncertainties introduced by renewable energy sources (RESs) can adversely impact cluster performance. In response to these challenges, this paper introduces a new distributed event-triggered consensus algorithm (DETC) to enhance the efficiency in handling the aforementioned issues. The proposed algorithm significantly reduces communication burdens, addressing resource usage concerns. The performance of this approach is evaluated through simulations of a cluster comprising four DC MGs, in each of which were PV and a hybrid Battery-Super capacitor in the MATLAB environment. The key findings indicate that the proposed DETC algorithm achieves commendable results in terms of voltage regulation, precise power sharing among sources, and a reduction in triggering instants. Based on these results, this method can be deemed as a good development in MGC management, providing a more efficient and reliable means of coordination, particularly in scenarios with dynamic loads and renewable energy integration. It is also a viable option for current microgrid systems, due to its ability to decrease communication loads while retaining excellent performance.

Information Freshness for Monitoring and Control over Wireless Networks

Brief Biography: Vishrant Tripathi obtained his PhD from the EECS department at MIT, working with Prof. Modiano at the Lab for Information and Decision Systems (LIDS). He is currently working on building efficient data center networks at Google. His research interests primarily lie in the optimization of resources in resource constrained networked systems. The main applications of his work are in multi-agent robotics, federated learning, edge computing, cloud infrastructure, and monitoring for IoT. More recently, he has also been working on software defined networking and next-generation wireless networks. In 2022, he won the Best Paper Runner Up Award at ACM MobiHoc.

Chaotic phase noise-like encryption based on geometric shaping for coherent data center interconnections.

The network traffic of data centers (DCs) has increased unprecedentedly with the rapid development of digital economy. However, the data transmission faces security threats in the distributed optical interconnection and intensive interaction of DC networks. In this paper, we propose a chaotic phase noise-like encryption algorithm using geometric shaping (GS) for coherent DC interconnections (DCIs). A GS constellation is used to improve transmission performance, and it is combined with coherent equalization algorithms to improve security performance. Then, a chaotic encryption is designed based on phase noise-like transformation (PNLT). The data are effectively scrambled, and the confusion level of phase can be increased. Finally, 216 Gb/s 8-quadrature amplitude modulation (8-QAM) encrypted data are successfully verified on a 240 km transmission link of DCIs. The results show that this scheme can achieve a bit error rate (BER) performance gain of 1.1 dB and provide a highly compatible solution for realizing security enhanced DCIs.

A Named Data Networking Based Fast Open Storage System Plugin for XRootD

This work presents the design and implementation of an Open Storage System plugin for XRootD, utilizing Named Data Networking (NDN). This represents a significant step in integrating NDN, a prominent future Internet architecture, with the established data management systems within CMS. We show that this integration enables XRootD to access data in a location transparent manner, reducing the complexity of data management and retrieval. Our approach includes the creation of the NDNc software library, which bridges the existing NDN C++ library with the high-performance NDN-DPDK data-forwarding system. This paper outlines the design of the plugin and preliminary results of data transfer tests using both internal and external 100 Gbps testbed.

MsDD: A novel NDN producer mobility support scheme based on multi-satellite data depot.

The Named Data Networking (NDN) is currently an important future network framework, and the mobility issue of producers within NDN is a primary challenge. However, in environments characterized by frequent producers mobility, traditional producer mobility support schemes still encounter issues such as excessive consumer delays and interest packet loss. With the development of The 6th generation communication technology (6G), integrating ground networks with satellites has emerged as a potential solution to address the aforementioned problems. In this paper, we propose an NDN producer mobility support scheme based on multi-satellite data depot, named MsDD. The proposed scheme proactively caches producer data into a data depot built on a low-earth orbit satellite constellation to minimize the impact of NDN producer mobility on network performance. We design data depot construction strategy, in-network caching strategy, and routing strategy based on forwarding hint to facilitate effective communication in satellite networks. Experimental results using ndnSIM demonstrate that compared with other existing schemes, MsDD can effectively shield the impact of producer mobility on consumer delay, delivery ratio, and signaling overhead, and MsDD has a clear advantage in terms of consumer delay and delivery ratio.

Job CPU Performance comparison based on MINIAOD reading options: Local versus remote

Job CPU Performance comparison based on MINIAOD reading options: Local versus remote

FMPTCP: Achieving High Bandwidth Utilization and Low Latency in Data Center Networks

FMPTCP: Achieving High Bandwidth Utilization and Low Latency in Data Center Networks

Fault-tolerant Hamiltonian cycle strategy for fast node fault diagnosis based on PMC in data center networks.

System-level fault diagnosis model, namely, the PMC model, detects fault nodes only through the mutual testing of nodes in the system without physical equipment. In order to achieve server nodes fault diagnosis in large-scale data center networks (DCNs), the traditional algorithm based on the PMC model cannot meet the characteristics of high diagnosability, high accuracy and high efficiency due to its inability to ensure that the test nodes are fault-free. This paper first proposed a fault-tolerant Hamiltonian cycle fault diagnosis (FHFD) algorithm, which tests nodes in the order of the Hamiltonian cycle to ensure that the test nodes are faultless. In order to improve testing efficiency, a hierarchical diagnosis mechanism was further proposed, which recursively divides high scale structures into a large number of low scale structures based on the recursive structure characteristics of DCNs. Additionally, we proved that $ 2(n-2){n^{k-1}} $ and $ (n-2){t_{n, k}}/{t_{n, 1}} $ faulty nodes could be detected for $ BCub{e_{n, k}} $ and $ DCel{l_{n, k}} $ within a limited time for the proposed diagnosis strategy. Simulation experiments have also shown that our proposed strategy has improved the diagnosability and test efficiency dramatically.

SAR: Receiver-driven Transport Protocol with Micro-burst Prediction in Data Center Networks

SAR: Receiver-driven Transport Protocol with Micro-burst Prediction in Data Center Networks

DemePro: DEcouple packet Marking from Enqueuing for multiple services with PROactive congestion control

Most of current Data Center Network (DCN) protocols leverage Explicit Congestion Notification (ECN) for congestion control to maintain low latency and high throughput. However, the majority of them only consider single-queue scenario in each switch port, making their performance inferior in multiple-service multiple-queue scenario. To this end, we propose DemePro, a DCN scheme for multiple-service multiple-queue scenario. In face of congestion signal, DemePro also leverages ECN for congestion notification, while decouples packet marking from enqueuing, in order to ensure fairness among multiple services. We also question the effectiveness of the congestion signal derived from the single threshold for port buffer queue length, via a set of experiments. Then, DemePro utilizes multiple thresholds for proactive congestion control, and packets are encapsulated to carry congestion extent information, in order to notify TCP senders for precise and fine-grained congestion control. Experiments show that DemePro has better performance than MQ-ECN [1] which is currently the best DCN scheme for multiple-service multiple-queue scenario, and DemePro can also well guarantee the fairness.

Lightweight Automatic ECN Tuning Based on Deep Reinforcement Learning With Ultra-Low Overhead in Datacenter Networks

Lightweight Automatic ECN Tuning Based on Deep Reinforcement Learning With Ultra-Low Overhead in Datacenter Networks

GTCC: A Game Theoretic Approach for Efficient Congestion Control in Datacenter Networks

GTCC: A Game Theoretic Approach for Efficient Congestion Control in Datacenter Networks

Next Generation Exabit/s Data Center Network Supporting All Optical Circuit and Packet Switching

Next Generation Exabit/s Data Center Network Supporting All Optical Circuit and Packet Switching

Dynamic Routing of Multiple QoS-Required Flows in Cloud-Edge Autonomous Multi-Domain Data Center Networks

Dynamic Routing of Multiple QoS-Required Flows in Cloud-Edge Autonomous Multi-Domain Data Center Networks

Latest advances in VCSEL technology for next-generation data center network [Invited

Latest advances in VCSEL technology for next-generation data center network [Invited

Novel Reliability Indicators From the Perspective of Data Center Networks

Novel Reliability Indicators From the Perspective of Data Center Networks