Data Center Networks Research Articles

Software-Defined Networking Challenges and Research Opportunities for Future Interest

Software Defined Network (SDN) has become one of the most preferred solutions for the management of large-scale complex networks. The network policies in the large-scale network are difficult to embed on entire network devices simultaneously, whereas in SDN these policies can be embedded on the top of the network. The SDN network is divided into two parts which are vertically integrated to form the entire network. Currently many aspects of the classical architecture of the Internet are etched in stone – a so-called ossification of the Internet – which has led to major obstacles in IPv6 deployment and difficulty in using IP multicast services. Yet, there exist many reasons to extend the Internet, e.g., for improving intra-domain and inter- domain routing for high availability of the network, providing end-to-end connectivity for users, and allowing dynamic management of network resources for new applications, such as data center, cloud computing, and network virtualization. To address these requirements, the next- generation architecture for the Future Internet has introduced the concept of Software-Defined Networking (SDN). At the core of this emerging paradigm is the separation and centralization of the control plane from the forwarding elements in the network as opposed to the distributed control plane of existing networks. With the advent of cloud computing, many ecosystem and business paradigms are encountering potential changes and may be able to eliminate their IT infrastructure maintenance processes. Real-time performance and high availability requirements have induced telecom networks to adopt the new concepts of the cloud model: software-defined networking (SDN) and network function virtualization (NFV). NFV introduces and deploys new network functions in an open and standardized IT environment, while SDN aims to transform the way network’s function. SDN and NFV are complementary technologies; they do not depend on each other. However, both concepts can be merged and have the potential to mitigate the challenges of legacy networks. In this paper, our aim is to describe the benefits of using SDN in a multitude of environments such as in data centers, data center networks, and Network as service offerings.

RP-DFC: Responsive probes and dynamic flow classification based load balancing in datacenter networks

RP-DFC: Responsive probes and dynamic flow classification based load balancing in datacenter networks

Predictive Maintenance for UWSNs: A Practical Implementation of Fault Detection Using NDN and Deep Learning

ABSTRACTUnderwater Wireless Sensor Networks (UWSNs) play a critical role in collecting data for various marine applications, but their deployment and maintenance in harsh underwater environments pose significant challenges. Building upon our prior conceptual work, this article presents a practical implementation of an AI‐driven fault detection system for UWSNs, a key component of a future predictive maintenance framework incorporating Named Data Networking (NDN). Using the “Underwater Sensor Dataset,” we implemented a feed‐forward neural network to classify sensor readings as healthy or faulty. The model achieved high accuracy (99.9%), precision (100%), recall (99.1%), and F1‐score (99.6%) on the test set, demonstrating its effectiveness in fault detection. Our findings highlight the feasibility and potential of AI‐based predictive maintenance in UWSNs, even with a simpler neural network model. The framework promises reduced downtime, minimized maintenance costs, and an extended lifespan for UWSNs.

Case Study: The Next Generation of Network Management - AI, Automation, and Security in a Connected World

Modern data center networks face unprecedented challenges in ensuring robust security due to the evolving complexity of cyber threats and the increasing sophistication of attack vectors. This study proposes comprehensive, multi-layered security architecture tailored for data center environments, integrating advanced technologies such as Next-Generation Firewalls (NGFWs), Zero Trust Architecture (ZTA), AI-driven anomaly detection, SQL-based policy management, and Neo4j knowledge graphs. The architecture leverages NGFWs for deep packet inspection and application-layer filtering, fortifying the network perimeter while enabling advanced threat detection. ZTA principles enforce least-privilege access, requiring continuous authentication and contextual validation for all users and devices. A relational database underpins security policy management, ensuring granular control and consistent enforcement across the network. Neo4j knowledge graphs offer a dynamic, graph-based visualization of the network topology, enabling real-time analysis of relationships and communication paths to uncover potential vulnerabilities, attack vectors, and insider threats. The core of the system’s intelligence lies in the integration of machine learning models, particularly Long Short-Term Memory (LSTM) networks, for anomaly detection and predictive analytics. By analyzing real-time network traffic data, the AI models autonomously detect unusual patterns indicative of security incidents, enabling proactive threat mitigation. The synergy between these components ensures a scalable and resilient security framework capable of addressing modern security challenges. This architecture is designed to automate key aspects of threat detection, incident response, and policy enforcement, significantly reducing operational overhead while improving response times. The result is a flexible and adaptive security solution that enhances visibility, control, and protection of critical data center resources. By combining these cutting-edge technologies, this proposed framework demonstrates its capability to provide a robust defense mechanism for data center networks, ensuring operational continuity and compliance with stringent security requirements. This paper highlights the system's technical components, demonstrates its functionality through a detailed use case, and underscores its effectiveness in securing complex, high-value network environments against evolving cyber threats.

Mitigating content poisoning attacks in named data networking: a survey of recent solutions, limitations, challenges and future research directions

Named Data Networking (NDN) is one of the capable applicants for the future Internet architecture, where communications focus on content rather than providing content. NDN implements Information-Centric Networking (ICN) with its unique node structure and significant characteristics such as built-in mobility support, multicast support, and efficient content distribution to end-users. It has several key features, including inherent security, that protect the content rather than the communication channel. Despite the good features that NDN provides, it is nonetheless vulnerable to a variety of attacks, the most critical of them is the Content Poisoning Attack (CPA). In this survey, the existing solutions presented for the prevention of CPA in the NDN paradigm have been critically analyzed. Furthermore, we also compared the suggested schemes based on latency, communication overhead, and security. In addition, we have also shown the possibility of other possible NDN attacks on the suggested schemes. Finally, we adds some open research challanges.

Nonblocking conditions for Clos fabrics with non-uniform switch radixes

Datacenter networks (DCNs) evolve over years and so comprise switches from different generations. Thus, each stage/layer of the Clos fabric may consist of switches with varying radixes (i.e., different port counts), leading to non-uniform stages. While optical circuit switches are increasingly deployed in DCNs to enhance transmission capacity and energy efficiency, the nonblocking condition, crucial for determining the performance of circuit-switched networks, has been established only for Clos fabrics with uniform stages. This study extends the nonblocking condition to Clos fabrics with non-uniform stages. To facilitate practicality, we formulate the condition using integer linear programming (ILP). Using our novel, to our knowledge, condition, we quantitatively demonstrate how much the nonblocking property is compromised under two practical scenarios, random link failures and network expansion, which would break network uniformity. In particular, we reveal that network expansion, common in DCN evolution, could significantly undermine the nonblocking property. Additionally, we assess the computational efficiency of our ILP formulation, which can successfully evaluate the nonblocking property of a large Clos fabric accommodating 32K terminals/uplinks in just 19 min.

Enhanced-Performance Tunable Sources for Fast AWGR-Based Optical Switching Data Center Networks

Enhanced-Performance Tunable Sources for Fast AWGR-Based Optical Switching Data Center Networks

Deep Learning Based Traffic Management in Knowledge Defined Network

In recent Artificial Intelligence developments, large datasets as knowledge are a prime requirement for analysis and prediction. To manage the knowledge of the network, the Data Center Network (DCN) has been considered a global data storage facility on edge servers and cloud servers. In recent research trends, knowledge-defined networking (KDN) architecture is considered, where the management plane works as the knowledge plane. The major network management task in the DCN is to control traffic congestion. To improve network management, i.e., optimized resource management, enhanced Quality of Service (QoS), we propose a path prediction technique by combining the convolution layer with the RNN deep learning model, i.e., Convolution-Long short-term memory network as Convolution-LSTM and the bi-directional long short-term memory (BiLSTM) network as Convolution-BiLSTM. The experimental results demonstrate that, in terms of many metrics, i.e., network latency, packet loss ratio, network throughput, and overhead, our proposed methodologies perform better than the existing works, i.e., OSPF, FlowDCN, modified discrete PSO, ANN, CNN, and LSTM-based routing approaches. The proposed approach improves the network throughput by approximately 30% and 12% as compared to existing CNN and LSTM-based routing approaches, respectively.

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

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

Reducing tail latency for multi-bottleneck in datacenter networks: A compound approach

The effectiveness of network congestion control fundamentally depends on the accuracy and granularity of congestion feedback. In datacenter networks, precise feedback is essential for achieving high performance. Most existing approaches use either Explicit Congestion Notification (ECN) or network delay (e.g., RTT) independently as congestion indicators. However, in multi-bottleneck networks, the limitations of these signals become more pronounced: ECN struggles with large cumulative end-to-end latency, while RTT lacks the precision needed to control queuing delays at individual hops. To address these challenges, we propose Cocktail, a simple yet effective transport protocol for datacenter networks that combines both ECN and RTT congestion signals to more effectively handle multi-bottleneck scenarios. By leveraging the ECN signal, Cocktail bounds per-hop queue lengths, enhancing its ability to control single-hop latency and prevent packet loss. Additionally, by estimating RTT, Cocktail effectively manages end-to-end delay, resulting in lower Flow Completion Time (FCT). Extensive experimental evaluations in Mininet demonstrate that Cocktail significantly reduces the average and 99th-percentile completion times for small flows by up to 20% and 29%, respectively, compared to current practices under production workloads.

A Receiver-Driven Named Data Networking (NDN) Congestion Control Method Based on Reinforcement Learning

Named data networking (NDN) is a novel networking paradigm characterized by in-network caching, receiver-driven communication, and multi-source, multi-path data retrieval, which poses new challenges for congestion control. Existing work has largely focused on receiver-driven mechanisms. Due to delays in obtaining network control information (timeouts, NACKs) within NDN, consumers are unable to access the network congestion status from this information in a timely manner. To address the issues above, this paper combines the Q-learning algorithm with the NDN architecture, proposing Q-NDN. In Q-NDN, consumers can dynamically adjust the congestion window (cwnd) through the real-time monitoring of network status, leveraging the Q-learning algorithm, achieving automatic congestion control for the NDN architecture. Additionally, this paper introduces content popularity-based traffic scheduling for multi-user scenarioswhich adjusts the transmission rates of content with different popularity levels to maintain a dynamic balance in the network. The experimental results show that Q-NDN can converge quickly, make full use of bandwidth resources, and keep the packet loss rate to 0 in the basic network topology. In competing network topologies, Q-NDN can rapidly address conflict issues, efficiently utilize bandwidth resources, and maintain a relatively low packet loss rate.

Interest Classification on Named Data Network Using the Supervised Learning Method

Named Data Network (NDN) is a next-generation network architecture that shifts the traditional data communications paradigm Unlike conventional networks that rely on IP addresses, NDN delivers content based on data names rather than specific locations. In NDN, consumers express their requests by sending interest packets containing content names. These names are then propagated through the network nodes, which forward them to the appropriate destinations. The forwarding strategy in an NDN network plays a crucial role in ensuring efficient data delivery. This strategy includes a set of rules that determine the next hop for each interest packet. These rules are designed to optimize the forwarding process, minimizing delays and improving network efficiency. However, if the forwarding strategy is implemented without accurately identifying the appropriate face (i.e., the network interface) to forward interests toward the producer or the nearest cache node, it can lead to significant delays and packet drops. This, in turn, negatively impacts Quality of Service (QoS) parameters and the overall performance of the NDN network. This study applies supervised learning to classify consumer-requested interests to overcome this issue. This technique leverages several related variables to accurately classify these interests. The outcomes of the conducted research demonstrated that raw data from the mini-NDN output can be processed and transformed into a usable dataset. This data is then utilized to train a classification model with supervised learning. In a scenario with 9 NDN nodes and varying numbers of interests, distributed both uniformly and according to Zipf's law, the Random Forest model performs effectively, achieving an accuracy rate of 86.2% with an error rate of 14.8%.

2-Edge Hamiltonian connectedness: Characterization and results in data center networks

A graph G is 2-edge Hamiltonian connected if for any edge set E⊆{uv:u,v∈V(G),u≠v} with |E|≤2, G∪E has a Hamiltonian cycle containing all edges of E, where G∪E is the graph obtained from G by including all edges of E. The problem of determining whether a graph is 2-edge Hamiltonian connected is challenging, as it is known to be NP-complete. This property is stronger than Hamiltonian connectedness, which indicates the existence of a Hamiltonian path between every pair of vertices in a graph. This paper first provides a characterization and a sufficiency for 2-edge Hamiltonian connectedness. Through this, we shed light on the fact that many well-known networks are 2-edge Hamiltonian connected, including BCube data center networks and some variations of hypercubes, and so on. Additionally, we demonstrate that DCell data center networks and Cartesian product graphs containing almost all generalized hypercubes are 2-edge Hamiltonian connected.

JamBIT: RL-based framework for disrupting adversarial information in battlefields

During battlefield operations, military radios (hereafter nodes) exchange information among various units using a mobile ad-hoc network (MANET) due to its infrastructure-less and self-healing capabilities. Adversarial cyberwarfare plays a crucial role in modern combat by disrupting communication between critical nodes (i.e., nodes mainly responsible for propagating important information) to gain dominance over the opposing side. However, determining critical nodes within a complex network is an NP-hard problem. This paper formulates a mathematical model to identify important links and their connected nodes, and presents JamBIT, a reinforcement learning-based framework with an encoder–decoder architecture, for efficiently detecting and jamming critical nodes. The encoder transforms network structures into embedding vectors, while the decoder assigns a score to the embedding vector with the highest reward. Our framework is trained and tested on custom-built MANET topologies using the Named Data Networking (NDN) protocol. JamBIT has been evaluated across various scales and weighting methods for both connected node and network dismantling problems. Our proposed method outperformed existing RL-based baselines, with a 24% performance gain for smaller topologies (50–100 nodes) and 8% for larger ones (400–500 nodes) in connected node problems, and a 7% gain for smaller topologies and 15% for larger ones in network dismantling problems.

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

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

Pushing the Performance Limits of Datacenter Networks With Fine-Grained Priority Assignment

Pushing the Performance Limits of Datacenter Networks With Fine-Grained Priority Assignment

A Decentralized Digital Watermarking Framework for Secure and Auditable Video Data in Smart Vehicular Networks

Thanks to the rapid advancements in Connected and Automated Vehicles (CAVs) and vehicular communication technologies, the concept of the Internet of Vehicles (IoVs) combined with Artificial Intelligence (AI) and big data promotes the vision of an Intelligent Transportation System (ITS). An ITS is critical in enhancing road safety, traffic efficiency, and the overall driving experience by enabling a comprehensive data exchange platform. However, the open and dynamic nature of IoV networks brings significant performance and security challenges to IoV data acquisition, storage, and usage. To comprehensively tackle these challenges, this paper proposes a Decentralized Digital Watermarking framework for smart Vehicular networks (D2WaVe). D2WaVe consists of two core components: FIAE-GAN, a novel feature-integrated and attention-enhanced robust image watermarking model based on a Generative Adversarial Network (GAN), and BloVA, a Blockchain-based Video frames Authentication scheme. By leveraging an encoder–noise–decoder framework, trained FIAE-GAN watermarking models can achieve the invisibility and robustness of watermarks that can be embedded in video frames to verify the authenticity of video data. BloVA ensures the integrity and auditability of IoV data in the storing and sharing stages. Experimental results based on a proof-of-concept prototype implementation validate the feasibility and effectiveness of our D2WaVe scheme for securing and auditing video data exchange in smart vehicular networks.

Towards fine-grained load balancing with dynamical flowlet timeout in datacenter networks

In modern datacenter networks (DCNs), load balancing mechanisms are widely deployed to enhance link utilization and alleviate congestion. Recently, a large number of load balancing algorithms have been proposed to spread traffic among the multiple parallel paths. The existing solutions make rerouting decisions for all flows once they experience congestion on a path. They are unable to distinguish between the flows that really need to be rerouted and the flows that potentially have negative effects due to rerouting, resulting in frequently ineffective rerouting. Fine-grained rerouting will also cause severe packet reordering, especially in asymmetric topology scenarios. To address the above issues, we present a fine-grained traffic-differentiated load balancing (TDLB) mechanism, which aims to distinguish flows that are necessarily to be rerouted and reroute traffic in fine-grained without packet reodering. Specifically, TDLB distinguishes the traffic that must be rerouted through the host pair information in the packet header, and selects an optimal path for rerouting. To prevent severe packet reodering caused by excessive path delay differences, TDLB dynamically adjusts the flowlet timeout to segment the traffic and select the optimal path for rerouting. The NS-2 simulation results show that TDLB effectively reduces tail latency and average flow completion time (FCT) for short flows by up to 49% and 46%, respectively, compared to the state-of-the-art load balancing schemes.

Optical circuit switched three-stage twisted-folded Clos-network design model guaranteeing admissible blocking probability

Some data center networks have already started to use optical circuit switching (OCS) with potential performance benefits, including high capacity, low latency, and energy efficiency. This paper addresses a switching network design to maximize the network radix, i.e., the number of terminals connected to the network under the condition that a specified number of identical switches with the size N×N and the maximum admissible blocking probability are given. Previous work presented a two-stage twisted and folded Clos network (TF-Clos) with a blocking probability guarantee for OCS, which has a larger network radix than TF-Clos with a strict-sense non-blocking condition. Expanding the number of stages allows for enhancing the network radix. This paper proposes a model designing an OCS three-stage TF-Clos structure with a blocking probability guarantee to increase the network radix compared to the two-stage TF-Clos. We formulate the problem of obtaining the network configuration that maximizes the network radix as an optimization problem. We conduct an algorithm based on an exhaustive search to obtain a feasible solution satisfying the constraints of the optimization problem. This algorithm identifies the structure with the largest network radix in non-increasing order to avoid unnecessary searches. Numerical results show that the proposed model achieves a larger network radix than the two-stage model.