Fig 8 - uploaded by Amaury Van Bemten
Content may be subject to copyright.
The four topologies considered in the evaluation are based on three different base topologies which can be scaled in two different directions.
Source publication
A variety of communication networks, such as industrial communication systems, have to provide strict delay guarantees to the carried flows. Fast and close to optimal quality of service (QoS) routing algorithms, e.g., delay-constrained least-cost (DCLC) routing algorithms, are required for routing flows in such networks with strict delay requiremen...
Contexts in source publication
Context 1
... first dimension of our evaluation framework, we define four topologies (shown in Fig. 8) based on three different base topologies. Although our survey is generic and all the algorithms can be applied to any CSP problem, we focus on industrial topologies where we expect centralized QoS routing to be extensively employed [133]. Nevertheless, the topologies we define are also common in and representative of data center, ...
Context 2
... and representative of data center, metro, grid, and enterprise networks. On the contrary, wide-area (star topology) networks are not covered, as strict centralized QoS routing in such environments is unlikely. All topologies can be scaled according to two scale parameters m and n that represent the size of the topology layout, as illustrated in Fig. 8 and defined in detail in the following for the four different topologies. The second and third dimensions of our evaluation framework correspond to varying the two scale parameters m and n from 4 to 13, thereby defining 100 different scalability levels. The four topologies are referred to as One Ring Bottleneck (ORB), Two Ring ...
Context 3
... first dimension of our evaluation framework, we define four topologies (shown in Fig. 8) based on three different base topologies. Although our survey is generic and all the algorithms can be applied to any CSP problem, we focus on industrial topologies where we expect centralized QoS routing to be extensively employed [133]. Nevertheless, the topologies we define are also common in and representative of data center, metro, grid, and enterprise networks. On the contrary, wide-area (star topology) networks are not covered, as strict centralized QoS routing in such environments is unlikely. All topologies can be scaled according to two scale parameters m and n that represent the size of the topology layout, as illustrated in Fig. 8 and defined in detail in the following for the four different topologies. The second and third dimensions of our evaluation framework correspond to varying the two scale parameters m and n from 4 to 13, thereby defining 100 different scalability levels. The four topologies are referred to as One Ring Bottleneck (ORB), Two Ring Bottleneck (TRB), Two Ring Random (TRR) and Grid Random ...
Context 4
... first dimension of our evaluation framework, we define four topologies (shown in Fig. 8) based on three different base topologies. Although our survey is generic and all the algorithms can be applied to any CSP problem, we focus on industrial topologies where we expect centralized QoS routing to be extensively employed [133]. Nevertheless, the topologies we define are also common in and representative of data center, metro, grid, and enterprise networks. On the contrary, wide-area (star topology) networks are not covered, as strict centralized QoS routing in such environments is unlikely. All topologies can be scaled according to two scale parameters m and n that represent the size of the topology layout, as illustrated in Fig. 8 and defined in detail in the following for the four different topologies. The second and third dimensions of our evaluation framework correspond to varying the two scale parameters m and n from 4 to 13, thereby defining 100 different scalability levels. The four topologies are referred to as One Ring Bottleneck (ORB), Two Ring Bottleneck (TRB), Two Ring Random (TRR) and Grid Random ...
Similar publications
![Fig. 4: [Not to Scale]. Depiction of a solution and its neighboring...](profile/Ajmal-Khan-3/publication/328924082/figure/fig3/AS:725445829984257@1549971172358/Not-to-Scale-Depiction-of-a-solution-and-its-neighboring-solution-The-left-part-of_Q320.jpg)
The throughput of a communication system depends on the data traffic load and the available capacity to support that load. In an unmanned aerial vehicle (UAV)-based communication system, the UAV position is one of the major factor affecting the capacity available to the flows (data sessions) being served. This paper proposes a centralized algorithm...
Citations
... How to schedule differentiated flow according to the specific QoS requirements is an important requirement for a wide Most existing approaches assume that the network state is relatively static during a period and model the QoS routing problem as optimization models [13]. However, the network topology will change constantly and the flows with different QoS requirements arrive and leave a network frequently. ...
Recently, 6G has attracted widespread attention from both academia and industry. 6G networks are expected to exhibit even more heterogeneity than 5G networks, and support various emerging scenarios and applications such as virtual and augmented reality (VR/AR), air/space/ground networks, and Internet of Things. Such massive heterogeneous devices pose huge challenges for network control and management. Recently, advances in artificial intelligence have brought to a new clan of networks, termed as self-driving 6G networks. It utilizes network telemetry, artificial intelligence, and DevOps to simplify networks and operations, helping network owners improve network quality, as well as increase efficiency. However, the current network is built on closed merchant switching ASICs and a homegrown management and control system. Self-driving an opaque system without really knowing and controlling what they do is a hazardous and fruitless endeavor. Fortunately, the network programmability technology opens the possibility for running self-driving algorithms over the whole network. In this paper, we design a full-dimensional programmability empowered self-driving 6G network architecture. We discuss how network programmability can promote the release of network intelligence, from the view of both verticality (control and data plane) and horizontality (end to end). Moreover, we design three use cases to demonstrate the feasibility and effectiveness of the proposed architecture.
... When such a link is broken at t 2 , the invocation will follow a different route (AS 0 , AS 2 , AS 5 , AS 4 ) and the QoS fluctuates accordingly. Using different routes to ensure QoS is crucial for various network settings and applications [12] has made routing estimation an important requirement [11]. ...
Dynamic quality of service (QoS) measurement is crucial for discovering services and developing online service systems. Collaborative filtering-based approaches perform dynamic QoS prediction by incorporating temporal information only but never consider the dynamic network environment and suffer from poor performance. Considering different service invocation routes directly reflect the dynamic environment and further lead to QoS fluctuations, we coin the problem of Dynamic QoS Prediction (DQP) with Intelligent Route Estimation (IRE) and propose a novel framework named IRE4DQP. Under the IRE4DQP framework, the dynamic environment is captured by Network Status Representation, and the IRE is modeled as a Markov decision process and implemented by a deep learning agent. After that, the DQP is achieved by a specific neural model with the estimated route as input. Through collaborative training with reinforcement and inverse reinforcement learning, eventually, based on the updated representations of the network status, IRE learns an optimal route policy that matches well with observed QoS values, and DQP achieves accurate predictions. Experimental results demonstrate that IRE4DQP outperforms SOTA methods on the accuracy of response-time prediction by 5.79-31.34% in MAE, by 1.29-20.18% in RMSE, and by 4.43-27.73% in NMAE and with a success rate of nearly 45% on finding routes.
... Early research on TE, which includes OSPF [20], ECMP, and QoS routing [21] along with their variants, focused on distributed IP and Multi-Protocol Label Switching networks. QoS routing, designed to address constrained shortest routing problems [22], helps find the shortest path that satisfies specific QoS demands such as bandwidth and delay. However, these algorithms may result in suboptimal decisions due to the lack of a global network view [2]. ...
... The primal problem (1) and the Lagrange dual problem (22) have strong duality according to the Slater's condition (i.e., strong duality holds when the primal problem is convex and strictly feasible). Therefore, a feasible iteration strategy is as follows. ...
... According to (24), x * jp is the solution to the maximum problem in the expression of ϕ(λ) in (22), which means ∂ϕ(λ) ∂λ ei = (C e − j∈Jt p∈P ...
The emergence of new applications brings multi-class traffic with diverse quality of service (QoS) requirements to wide area networks (WANs), motivating research in traffic engineering (TE). In recent years, novel centralized and hierarchical TE schemes have used heuristic or machine learning techniques to orchestrate resources in closed systems such as datacenter networks. However, these schemes suffer from long delivery delays and high control overhead when applied to general WANs. To provide low-delay services, this paper proposes an asynchronous multi-class traffic management (AMTM) scheme. We first establish an asynchronous TE paradigm in which distributed nodes locally perform low-complexity and low-delay traffic control based on link prices, and the TE server updates link prices to eliminate decision conflicts between edge nodes. By modeling the asynchronous TE paradigm as a control system with non-negligible control loop delay, we find that the traditional pricing strategy cannot simultaneously achieve a low packet loss rate and a low flow delivery delay. To address this issue, we propose a new pricing strategy based on the observations of virtual queues in intermediate nodes. We also present a system design and related algorithms that utilize a dynamic step size mechanism of link price update. Simulation results show that AMTM can effectively reduce the end-to-end flow delivery delay.
... Hence, meeting the QoS requirements is the key to evaluating the effectiveness of the routing strategy. Moreover, depending on the QoS requirements, a different problem can be defined to find the optimal routing path [51], where the most common are the following ones: ...
... Different surveys [51], [55], [56] compile and study other SP algorithms, such as the Bellman-Ford and the Floyd-Warhsall ones. These are similar to Dijkstra algorithm. ...
... It works by using the Bellman-Ford algorithm to compute a transformation of the input graph that removes all negative weights, allowing Dijkstra's algorithm to be used on the transformed graph. Finally, as indicated at [51], the A* algorithm was proposed by Hart et al. for finding a singledestination SP by introducing a so-called guess function. At each node, this guess function provides an estimation for the cost of the SP from this node to the destination node. ...
The growth of network traffic and the rise of new network applications having heterogeneous requirements are stressing the telecommunication infrastructure and pushing network management to undergo profound changes. Network management is becoming a core research area to push the network and its performance to the limits, as it aims at applying dynamic changes across the network nodes to fit the requirements of each specific network traffic or application. Here, solutions and frameworks based on software-defined networking (SDN) and network function virtualization (NFV) facilitate the monitorization and control of both the network infrastructure and the network services running on top of it. This article identifies and analyzes different implemented solutions to perform experiments on network management. In this context, an innovative experimental testbed is described and implemented to allow experimentation. A predictive path routing algorithm is later proposed and tested by designing experiments with specific network topologies and configurations deployed through the testbed. The algorithm exploits predictions on network latency to change the routing rules. Finally, the article identifies the open challenges and missing functions to achieve next-generation network management.
... This decision-maker needs to master the configuration information of the entire network and determine the reconfiguration strategy and specific algorithm based on the fault type, the current network configuration information, and the resource distribution. Software-defined networks (SDN) are an architecture solution that meets the above requirements, achieving rapid response and deployment of network-wide reconfiguration by deploying the decision-maker in the control plane [28]. The fundamental principle of SDN is to separate the control plane from the data plane responsible for traffic forwarding in traditional networks. ...
In avionics system networking, mixed‐critical networks are commonly used to integrate system applications with different levels of guarantees for various traffic classes on a shared resource platform, effectively addressing issues such as cabling, cost, weight, volume, power consumption, etc. However, current fault recovery and reconfiguration techniques mainly focus on high‐priority time‐triggered (TT) traffic, often neglecting the impact of TT traffic reconfiguration on low‐priority traffic. This paper proposes an avionics mixed‐critical networking approach based on the IEEE 802.1CB protocol, which is shown to be more effective for the post‐failure recovery of Rate‐Constrained (RC) traffic (a typical type of low‐priority traffic in TT networks). Furthermore, an elastic reconfiguration algorithm is presented to ensure quality of service for RC traffic during and after reconfiguration, preventing arbitrary application and traffic shutdown triggered by TT traffic reconfiguration. Simulations demonstrate that the proposed approach and algorithm can significantly improve the post‐failure recovery ratio of RC traffic, as well as guarantee quality of service and network throughput of RC traffic during and after reconfiguration to a certain extent.
... Note that the greedy Algorithm 1 needs to go through only two iterations (one iteration for z and one for w) of a shortestpath-first algorithm to find the shortest path, resulting in a time complexity of O(V 2 ) [55]. This article has been accepted for publication in IEEE Transactions on Network and Service Management. ...
Industrial Internet of Things (IIoT) applications have diverse network session requirements. Certain critical applications, such as emergency alert relays, as well as industrial floor evacuation and surveillance systems, require fresh updates that can maintain the most recently delivered packets. This requires high reconfigurability to an extent where the system can measure the impact of an event and adapt the network accordingly. Recent research has demonstrated that network failures can undermine the sustainability of Industry 4.0 or Industrial IoT in general. In this paper, we design an intelligent Federated learning based Time-Sensitive Networking (Fed-TSN) controller framework to optimize the failure recovery. In industrial IoT scenarios, such as emergency evacuations on factory floors due to natural disasters, there can be multiple link failures with no disjoint paths which require a sustainable recovery solution. Accordingly, we consider multiple simultaneous link failures, both for networks with and without disjoint paths. The typically probabilistic network failures on a factory floor call for designing a mechanism that can search for routes with minimum joint failure probability (JFP). We formulate the JFP minimization problem as a non-linear integer program. We design a Software Defined Networking (SDN) controller that runs an application to produce near-optimal solutions for providing enhanced sustainability in a wide range of Industry 4.0 scenarios. We employ this non-linear integer program solution as input to our intelligent Fed-TSN fault recovery strategy that predicts the migration location based on the changes in the TSN gate schedule. We conduct simulations to quantify the improvements achieved with Fed-TSN compared to state-of-the-art approaches.
... Early research on TE, which includes OSPF [18], ECMP, and QoS routing [19] along with their variants, focused on distributed IP and Multi-Protocol Label Switching networks. QoS routing, designed to address constrained shortest routing problems [20], helps find the shortest path that satisfies specific QoS demands such as bandwidth and delay. However, these algorithms may result in suboptimal decisions due to the lack of a global network view [2]. ...
The emergence of new applications brings multi-class traffic with diverse quality of service (QoS) demands in wide area networks (WANs), which motivates the research in traffic engineering (TE). In recent years, novel centralized TE schemes have employed heuristic or machine-learning techniques to orchestrate resources in closed systems, such as datacenter networks. However, these schemes suffer long delivery delay and high control overhead when applied to general WANs. Semi-centralized TE schemes have been proposed to address these drawbacks, providing lower delay and control overhead. Despite this, they suffer performance degradation dealing with volatile traffic. To provide low-delay service and keep high network utility, we propose an asynchronous multi-class traffic management scheme, AMTM. We first establish an asynchronous TE paradigm, in which distributed nodes instantly make traffic control decisions based on link prices. To manage varying traffic and control delivery time, we propose state-based iteration strategies of link pricing under different scenarios and investigate their convergence. Furthermore, we present a system design and corresponding algorithms. Simulation results indicate that AMTM outperforms existing schemes in terms of both delay reduction and scalability improvement. In addition, AMTM outperforms the semi-centralized scheme with 12-20$\%$ more network utility and achieves 2-7$\%$ less network utility compared to the theoretical optimum.
... In contrast, this paper focused on SDN PSA, classified based on path selection criteria. Tomovic et al. [40] and Guck [41] compared the QoS routing in large-scale SDN with a focus on bandwidth and delay. Guck's study focused on unicast communication and considered only the algorithms that find a single path. ...
... QoS satisfaction deals with the ability of a network to consistently adhere to the performance expected by an application in terms of delay, jitter, bandwidth, throughput, and loss. QoS routing is one of the most critical components of a network management framework [41]. Ensuring it in a path is challenging due to network dynamics. ...
Software Defined Networking (SDN) introduced network management flexibility that eludes traditional network architecture. Nevertheless, the pervasive demand for various cloud computing services with different levels of Quality of Service requirements in our contemporary world made network service provisioning challenging. One of these challenges is path selection (PS) for routing heterogeneous traffic with end-to-end quality of service support specific to each traffic class. The challenge had gotten the research community's attention to the extent that many PSAs were proposed. However, a gap still exists that calls for further study. This paper reviews the existing PSA and the Baseline Shortest Path Algorithms (BSPA) upon which many relevant PSA(s) are built to help identify these gaps. The paper categorizes the PSAs into four, based on their path selection criteria, (1) PSAs that use static or dynamic link quality to guide PSD, (2) PSAs that consider the criticality of switch in terms of an update operation, FlowTable limitation or port capacity to guide PSD, (3) PSAs that consider flow variabilities to guide PSD and (4) The PSAs that use ML optimization in their PSD. We then reviewed and compared the techniques' design in each category against the identified SDN PSA design objectives, solution approach, BSPA, and validation approaches. Finally, the paper recommends directions for further research.
... There are several studies carried out to evaluate the performance of SDN [24], [25], [26]. Also, several studies [27], [28], [29], [30], [31] have produced solutions for data traffic offloading systems. ...
The approach of data traffic offloading methodologies is likely to improve the quality of mobile service to address the issue of insufficient bandwidth due to the rapid growth of cellular data traffic. To measure the real‐time performance of Software‐defined networking (SDN) based offloading systems, computing the response time is essential to consider. In this study, we develop a computation model to estimate the response time of the SDN‐based data traffic offloading system (SDN‐TOS) to predict the efficiency of system performances accurately. The values related to the process of Mininet emulator were collected from a mobile communication company through a third‐party broker based in Sri Lanka. Further analysis is considered to perform the comparison between the proposed model and the Cloud Service Providers (CSP) approach. The CSP approach considers only one network to estimate the response time; in contrast, our model perceives the response time of the SDN controller and both Long‐Term Evolution (LTE), and Wi‐Fi in the offloading process. Hence, our computation model generates high accurate value for the required response time of SDN‐TOS. The essential parameters that directly affect the offloading task such as computation capability and uplink data rate are observed through the comparison between two different service providers. The computation capability and uplink data rate of data traffic offloading processes are involved in a significant role in real‐time decision making for data and mobile communication services. Our analysis exhibits the effectiveness of a comprehensive computation model and identifies the most appropriate parameters to enhance the performance of SDN‐TOS in the mobile and data communication industries.
... In addition to reviewing QoS routing algorithms for the past few years, they also present a novel 4dimensional evaluation framework for QoS routing algorithsm. This may contribute to more scientific and reasonable evaluation on performance of QoS optimization algorithms for unicast routing [7]. Karakus et al. summarize QoS optimization schemes in OpenFlow protocol based SDNs, involving varieties of technical solutions aimed at QoS optimization (e.g., a multimedia stream transmission mechanism, inter-domain routing, a resource protection mechanism, queue management and scheduling algorithms) [8]. ...
Nowadays, ocean network gradually becomes more and more important for communication among network entities such as maritime and sea-crossing users. However, ocean network is highly dynamic because the communication links are composed of satellite and microwave links, which could be easily influenced by the environment such as local climate. Thus, network transmission in ocean networks faces great challenges, including low reliability and low efficiency. In this paper, we propose a smart ocean network architecture, where we use Software Defined Network (SDN) to perform unified management of the network, and Segment Routing (SR) to control data forwarding paths. In this way, we can control network flows and optimize network routing among diverse network entities in an ocean network. However, there are many QoS guaranteed applications in oean networks, such as remote control which requires lower delay. To guarantee the performance for such applications, we further propose QoS routing algorithms based on Fuzzy-Lagrange for the smart ocean network architecture, where the optimization objective is to ensure service quality provided to users. According to experimental results, it is proved that, in comparison with the benchmark algorithms, the Fuzzy-Lagrange (FuzLag) algorithm proposed based on link fuzzification and Lagrangian Relaxation can improve the performance by 23% at most.
