QoS Requirements Research Articles

DVFaaS: Leveraging DVFS for FaaS Workflows

Serverless computing's promise for fine-grained and modular functions that offer unprecedented elasticity, amplifies the challenge for efficient resource management. Functions form chains to compose the services an application provides. To match the practicality of the conventional deployment types, serverless chains require performance regulation to achieve QoS requirements. At the infrastructure level, a compelling way to regulate the performance of workloads while preserving power efficiency is through Dynamic Voltage and Frequency Scaling. However, existing generic frequency governors are rendered inadequate for QoS-aware power management in serverless deployments. In this paper, we propose <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DVFaaS</i> , a per-core DVFS framework that utilizes control systems theory to assign <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">just-enough</i> frequency for the purpose of addressing the QoS requirements on serverless workflows comprising unseen functions. <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DVFaaS</i> exploits the intermittent nature of serverless workflows, which enables staged control on distinguishable functions, which jointly contribute to the end-to-end latency. Our results show that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">DVFaaS</i> considerably outperforms related work, reducing power consumption by up to 22%, with 2x fewer QoS violations.

Challenges and Opportunities in QoS Enhancement for Internet-Based Multimedia Communication

Advancements in communication technologies have fueled the rapid expansion of the multimedia content delivery market, encompassing various media types such as images, audio, and video. Personal communication, computing, broadcasting, and entertainment heavily rely on multimedia content, which is transmitted through diverse communication and network technologies to reach a wide range of devices. QoS requirements for multimedia services are determined by parameters including bandwidth flexibility, low end-to-end delay, minimal delay variation, and acceptable error rates. These parameters are interconnected and vary depending on the type of multimedia application. However, the QoS needs of multimedia communications cannot be fully met by the internet's basic infrastructure. Mechanisms for ensuring QoS guarantees over the Internet have been developed through extensive research and development. This study examines developments, difficulties, and strategies for improving the transmission of multimedia content over the Internet, providing an overview of the methodologies and strategies used to satisfy QoS criteria in multimedia communications.

Preference-Aware User Access Control Policy in Internet of Things.

There are multiple types of services in the Internet of Things, and existing access control methods do not consider situations wherein the same types of services have multiple access options. In order to ensure the QoS quality of user access and realize the reasonable utilization of Internet of Things network resources, it is necessary to consider the characteristics of different services to design applicable access control strategies. In this paper, a preference-aware user access control strategy in slices is proposed, which can increase the number of users in the system while balancing slice resource utilization. First, we establish the user QoS model and slice QoS index range according to the delay, rate and reliability requirements, and we select users with multiple access options. Secondly, a user preference matrix is established according to the user QoS requirements and the slice QoS index range. Finally, a preference matrix of the slice is built according to the optimization objective, and access control decisions are made for users through the resource utilization state of the slice and the preference matrix. The verification results show that the proposed strategy not only balances slice resource utilization but also increases the number of users who can access the system.

Two-stage hybrid genetic algorithm for robot cloud service selection

Robot cloud service platform is a combination of cloud computing and robotics, providing intelligent cloud services for many robots. However, to select a cloud service that satisfys the robot’s requirements from the massive services with different QoS indicator in the cloud platform is an NP hard problem. In this paper, based on the cost model between the cloud platform, cloud services and cloud service robotics, we propose a two-stage service selection strategy, namely, candidate services selection stage according to the specific QoS requirements of service robots and final cost optimization stage. Additionally, with respect to optimizing the final cost for the model, we propose a Dynamic Vector Hybrid Genetic Algorithm (DVHGA) that is integrated with local and global search process as well as a three-phase parameter updating policy. Specifically, inspired by momentum optimization in deep learning, dynamic vector is integrated with DVHGA to modify the weights of QoS and ensure the reasonable allocation of resources. Moreover, we suggest a linear evaluation method for the service robots and the cloud platform concerning time and final cost at the same time, which could be expected to be used in the real application environment. Finally, the empirical results demonstrate that the proposed DVHGA outperforms other benchmark algorithms, i.e., DABC, ESWOA, GA, PGA and GA-PSO, in convergence rate, total final cost and evaluation score.

Cooperative UAV Resource Allocation and Task Offloading in Hierarchical Aerial Computing Systems: A MAPPO-Based Approach

This paper investigates a hierarchical aerial computing system, where both HAPs and UAVs provision computation services for ground devices (GDs). Different from the existing works which ignored UAV task offloading to HAPs and suffered long transmission delay between HAPs and GDs, in our system, UAVs are responsible for collecting the tasks generated by GDs. Considering limited resources and constrained coverage, UAVs need to cooperatively allocate their resources (including spectrum, caching, and computing) to GDs. After collecting GD tasks, UAVs are allowed to offload part of these tasks to the HAP, in order to minimize task processing delay and then better satisfy GD delay requirement. Our objective is to maximize the amount of computed tasks while satisfying tasks heterogeneous QoS requirements through the joint optimization of UAV resource allocation and task offloading. To this end, a joint optimization problem is first formulated as a partially observable Markov decision process (POMDP) under the constraints of available resources, UAV energy, and collision avoidance. Then, we design a multi-agent proximal policy optimization (MAPPO)-based algorithm to solve the optimization problem. By introducing the centralized training with decentralized execution framework, UAVs acting as agents can cooperatively make decisions on ground devices association, resource allocation, and task offloading according to their local observations. In addition, state normalization and action mask are also adopted to improve training efficiency. Experiment results verify the efficiency of the proposed algorithm and the system performance is also analyzed by the numerical results.

The new trends in networking: IOT, 5G and cognitive radio

The domain of networking and wireless communication is rapidly expanding to the point where traditional legacy networks research is nearly complete, and they are supposed to be an elder. IoT, the Internet of Things, and the 5G radio communications are the newest networking advancements. These two technologies fulfil all of the researchers' goals of improving people's quality of life. IoT is a framework that allows everyday devices to become smarter, processing to become more intelligent, and communication to become more relevant. The Internet of Things (IoT) has already made huge progress as a ubiquitous solution platform for the linked world. We also discussed the Radio Cognitive RC, which allows people to connect to networks as secondary users at no cost while the service provided is not as good as that provided to primary users who pay a license to use the network. RC can provide a solution for best-effort IoT applications that do not require constant connectivity and high QoS requirements. This study describes all of these technologies and concludes with a recommendation for using IoT and RC in our further works.

Trajectory Planning of UAV-Enabled Data Uploading for Large-Scale Dynamic Networks: A Trend Prediction Based Learning Approach

The unmanned aerial vehicle (UAV) enabled communication technology is regarded as an efficient and effective solution to provide emergency data uploading for some special cases where existing cellular infrastructures cannot provide reliable services to large-scale ground users (GUs). In the face of large-scale dynamic networks where the number and location of GUs change all the time, to maximize the throughput of data upload tasks under the premise of meeting the requirements of fairness, QoS and energy consumption, we propose a UAV trajectory planning approach based on deep reinforcement learning in combination with the prediction of GUs' movement trend in future time slots. Simulation results validate that our proposed approach converges more quickly and achieves better performance in dynamic networks.

ViTaLS—A Novel Link-Layer Scheduling Framework for Tactile Internet Over Wi-Fi

The pioneering field of Tactile Internet (TI) will enable the transfer of human skills over long distances through haptic feedback. Realizing this demands a roundtrip latency of sub-5 ms. In this work, we investigate the capability of Wi-Fi 6 and existing TI scheduling/multiplexing schemes in meeting this stringent latency constraint. Taking the concrete example of the state-of-the-art Video-Haptic multiplexer (VH-multiplexer), we highlight the pitfalls of relying on the existing Wi-Fi 6 systems for TI communication. To circumvent this, we propose Video-Tactile Latency Scheduler (ViTaLS) -a novel link layer framework for tuning the video-tactile frame transmissions to suit their heterogeneous QoS requirements. We present a mathematical model to characterize the packet transmission duration of Vi-TaLS. Using a custom simulator, we validate our model and measure the objective performance improvement of ViTaLS over VH-multiplexer. We also present ViTaLS-optimal -a variant of ViTaLS, for further reducing the tactile latency. Objectively, we show that ViTaLS-optimal yields a latency improvement of up to 82%. Based on experiments conducted on a real TI testbed, we subjectively demonstrate that ViTaLS-optimal outperforms the VH-multiplexer.

PARS: A Priority-Aware Resource Sharing Scheme for Services With Differential QoS Requirements in B5G Edge Network

Next-generation networks are expected to address scenarios demanding varying QoS, such as higher bandwidth, lower latency, and reliable data delivery. A multitude of B5G applications like AR/VR/XR and health-care needs differential QoS requirements and higher reliability of internet services. In this context, we develop a novel approach, PARS, for resource provisioning among slices for facilitating efficient resource utilization in an edge network. Further, we propose an adaptive SFC-RA algorithm to maximize user admission under computation and bandwidth resource constraints while satisfying SLA requirements of health-care services. Moreover, we show efficacy of our solution by comparing it with existing state-of-the-art methods.

Exploring the potential of dynamic quality of service cloud-based network slicing in 5G and next generation virtualized networks: A simulation-based study

With global mobile traffic doubling every 18 months along with the massive increase in unique data users across digital platforms during the Covid-19 pandemic, the world's demand for high-speed connectivity continues to increase every second. The Covid-19 pandemic has transformed the global workforce with remote working arrangements, embracing new platforms, subsequently increasing the requirement for reliable, quality, ubiquitous high speed data connectedness everywhere. 5th Generation (5G) mobile communication technology provides an effective solution, modernizing network infrastructure, efficiently providing reliable high bandwidth capacity for massive data growth and ultra-low latency, making delays virtually impossible to perceive. This means elevating reality even further with fiber-like speeds for everyone, everywhere. 5G network slicing makes it possible for services to own dedicated portions of the 5G network with guaranteed performance for their particular need. These virtual networks allow for a service like IoT, at industrial scale, making society more sustainable and increasing resource efficiency. Our paper proposes a model for 5G network slice creation that meets industry verticals service QoS requirements which are critical to service adoption. A software emulation is utilized to guide the design and evaluation of our proposed model. Simulated results show QoS guarantees are met for the reviewed service requirements in varying network conditions.

User QoS-Based Optimized Handover Algorithm for Wireless Networks

Due to the development of wireless network technology, various applications relying on good network quality are widely used on mobile devices. Taking the commonly used video streaming service as an example, a network with high throughput and low packet loss rate can meet the service requirements. When the moving distance of the mobile device is greater than the signal coverage of the AP, it will trigger the handover process to connect to another AP, and cause the network to disconnect and reconnect instantaneously. However, frequently triggering the handover procedure will cause a significant drop in network performance and affect the operation of application services. In order to solve this problem, this paper proposes the OHA and OHAQR. The OHA considers whether the signal quality is good or bad, and uses the corresponding HM method to solve the problem of frequent handover procedures. The OHAQR integrates the QoS requirements of the throughput and packet loss rate into the OHA with the Q-handover score, to provide high-performance handover services with QoS. Our experimental results show that the OHA and OHAQR have 13 and 15 handovers in a high-density scenario, respectively, and are better than the other two methods. The actual throughput and packet loss rate of the OHAQR are 123 Mbps and 5%, respectively, and the network performance is better than that of other methods. The proposed method shows excellent performance in ensuring the network QoS requirements and reducing the number of handover procedures.

Minimum cost spectrum allocation with QoS guarantees in multi-interface multi-hop dynamic spectrum access networks

Spectrum Sensing as a Service (SSaS) is emerging as the enabling business model for efficient spectrum sharing in many recent applications. In such model, sensing infrastructure provides dynamic spectrum access networks, also known as cognitive radio networks, with information about availability or unavailability of given spectrum bands. In return for such information, the service provider imposes money costs on the users of the dynamic spectrum access network or its operator. In this paper, we address the problem of spectrum allocation along a route in multi-interface dynamic spectrum access networks under the SSaS model. The objective is to allocate spectrum channels to interfaces of nodes along the route so that the total sensing cost imposed by SSaS provider is minimized and the quality of links meet a predetermined QoS requirement, specifically a data rate requirement. The problem is formulated as an Integer Linear Program (ILP) to obtain the optimal solution. Given the intractable complexity of ILPs, a sub-optimal algorithm is proposed. The accuracy of the proposed algorithm is validated via extensive experimentation.

MJOA-MU: End-to-edge collaborative computation for DNN inference based on model uploading

As an emerging computing paradigm, edge computing can assist user equipments (UEs) in executing computation-intensive deep neural network (DNN) inference tasks, thereby satisfying the stringent QoS requirement and relieving the burden of UEs. Due to the customizability of DNN models and limited capacity of the edge server, it is more realistic to upload DNN models on demand during end-to-edge co-inference, instead of deploying all DNN models at the edge server in advance. Existing works adopt the serial model uploading manner that uploads subsequent DNN layers only after antecedent DNN layers finish execution, inevitably prolonging the DNN execution latency. To this end, we innovatively design a parallel-efficient model uploading mechanism that allows subsequent DNN layers to be uploaded simultaneously when executing antecedent DNN layers, so as to efficiently mitigate the performance drop caused by model uploading. On this basis, we propose a Multi-UE Joint Optimization Algorithm based on Model Uploading (MJOA-MU) to optimize DNN partitioning and resource allocation for heterogeneous UEs. Specifically, MJOA-MU includes a Pruned Binary Tree based DNN Partitioning (PBT-DP) sub-algorithm to efficiently make the near-optimal partitioning decision for chain and non-chain models based on the long-term influence between DNN layers, and an Asynchronous Resource Allocation (ARA) sub-algorithm to allocate computation and communication resources for UEs by quantifying the inner- and inter-association, so as to match with individual demand and resource budget. Extensive simulation results demonstrate that MJOA-MU outperforms the state-of-the-art in terms of the DNN execution latency, and specifically achieves up to 64.5% reduction.

Raptor-IRSA Grant-free Random Access protocol for smart grids applications

This paper deals with the reliability of random access (RA) protocols for massive wireless smart grid communication (m-SGC). We propose and analyze an improved grant-free RA (GF-RA) protocol for critical SG applications under strict QoS m-SGC requirements. At first, we discuss the main features of the SG neighborhood area network (NAN) architecture. We explore the main features of low-rate machine-type wireless networks, and also we describe a technology characterization of wireless neighborhood area networks (WNAN) in medium-range coverage applications. We propose a new-improved irregular repetition slotted ALOHA, combining Raptor codes and irregular ALOHA, namely RapIRSA random access protocol, to better respond to critical high-reliability QoS requirements under a 5G network perspective. Then, we compare and comprehensively analyze the proposed RapIRSA protocol with two existing RA protocols, the IRSA protocol, and the classical Slotted Aloha. Finally, We summarize the potential challenges in implementing the proposed RA protocol for SG critical applications considering a massive number of smart sensors (SS).

A hybrid Square Shaped Adaptive Neuro-Fuzzy Interference with M-PeSOA model for optimal path selection

In Vehicular cloud computing (VCC), at the time of congestion, to deal with traffic, the vehicles’ underutilized resources are shared; subsequently, these resources aren’t constrained to computing power, storage, along with internet connectivity. Nevertheless, owing to the vehicular network characteristics, attaining the QoS requirements search together with the allocation of resources in the Vehicular Cloud (VC) has turned into a complicated task. An intelligent Square Shaped (SS)-Adaptive Neuro-Fuzzy Interference (SS-ANFIS) methodology for Resource Scheduling (RS) in addition to Mean-centered Penguins Search Optimization Algorithm (M-PeSOA) for Optimal Path Selection (OPS) in the VC is proposed here for efficient resource allocation. (a) Feature extraction, (b) Vehicles clustering, (c) OPS, (d) Resource information extraction, and (e) RS included in the proposed methodology. First, the vehicular network is initialized following that the vehicle features are extracted. Next, Cluster Heads (CHs) are generated regarding which vehicles are clustered; subsequently, the multi-paths are generated. After that, by employing the M-PeSOA, the OPS procedure is conducted; thus, the VC’s resource information is extracted aimed at scheduling the resources efficiently. Lastly, by employing the SS-ANFIS, vehicles are scheduled in the optimal paths. The proposed resource allocation system’s performance is assessed, and the experiential outcomes are analogized using the sumo tool and java platform.

Leveraging Deep Reinforcement Learning With Attention Mechanism for Virtual Network Function Placement and Routing

The efficacy of Network Function Virtualization (NFV) depends critically on (1) where the virtual network functions (VNFs) are placed and (2) how the traffic is routed. Unfortunately, these aspects are not easily optimized, especially under time-varying network states with different QoS requirements. Given the importance of NFV, many approaches have been proposed to solve the VNF placement and Service Function Chaining (SFC) routing problem. However, those prior approaches mainly assume that the network state is static and known, disregarding dynamic network variations. To bridge that gap, we leverage Markov Decision Process (MDP) to model the dynamic network state transitions. To jointly minimize the delay and cost of NFV providers and maximize the revenue, we first devise a customized Deep Reinforcement Learning (DRL) algorithm for the VNF placement problem. The algorithm uses the attention mechanism to ascertain smooth network behavior within the general framework of network utility maximization (NUM). We then propose attention mechanism-based DRL algorithm for the SFC routing problem, which is to find the path to deliver traffic for the VNFs placed on different nodes. The simulation results show that our proposed algorithms outperform the state-of-the-art algorithms in terms of network utility, delay, cost, and acceptance ratio.

Radio Resource Allocation for 5G Network Using Deep Reinforcement Learning

Abstract: Resource allocation is a critical task in 5Gnetworks that determines how network resources are assigned to different devices and services. Traditional methods rely on predefined rules or heuristics, which may not always be optimal. Deep reinforcement learning (DRL)is a promising approach for radio resource allocation in 5Gnetworks as it can learn to optimize resource allocation based on feedback from the network. In DRL, an agent learns to make decisions based on rewards and penalties received from the environment. In radio resource allocation, the agent would learn to allocate resources, such as frequency bands and power levels, to different devices and services to maximize some performance metric, such asthroughput or energy efficiency. The main challenge in applying DRL to radio resource allocation is designing an appropriate reward function that incentivizes the agent to improve the performance metric while avoiding undesirable behavior. Additionally, the radio resource allocation problem is complex, requiring the agent to consider many variables and constraints, such as channel conditions, interference, and QoS requirements. To address this, researchers have proposed various techniques such as hierarchical RL, multi-agent RL, and curriculum learning. Despite the challenges, DRL has shown promising results inradio resource allocation for 5G networks. It has outperformed traditional methods in some scenarios, especially when network conditions are dynamic and unpredictable. However, further research is necessary to explore the scalability and robustness of DRL-based approaches in practical 5G networks. In this method we suggest an algorithm for voice and data carriers in sub-6 GHz and millimeter wave (mmWave) frequencies respectively. The mmWave ranges between 30GHz to 300GHz

Energy‐efficient task scheduling and resource management in a cloud environment using optimized hybrid technology

AbstractCloud computing is a significant platform emerging for IT enterprises, business applications, and mobile computing. It is a dynamic environment where efficiently and properly allocating resources in such an environment becomes a tedious task. Most existing approaches cannot guarantee energy‐efficient resource management due to the existence of time‐dependent tasks. Cloud resources consume a huge amount of energy, which may reduce the Makespan of the entire network. The process of appropriate task scheduling may satisfy the user's requirement. Reducing energy consumption by satisfying the user's QoS requirements is essential to ensure each user's service level agreement (SLA). Therefore, to achieve an energy‐efficient and unique task scheduling, a technique for order preference by similarity to ideal solution (TOPSIS) based unique ranking (uRank‐TOPSIS) and a Hybrid state‐action‐reward‐state‐action (SARSA), Reinforcement Learning with Black widow Algorithm (HSRLBA) are proposed in this paper. Here, the task scheduling process is carried out using the uRank‐TOPSIS process by producing a set of unique weights and ranking the alternatives. Also, HSRLBA is used to perform resource allocation, which uses the Black widow Algorithm (BWA) to rapidly converge parallel agents in the SARSA model of RL. The Cloudsim simulator is used for experiments. The efficiency of the proposed framework is measured in terms of Makespan, Task Completion Ratio (TCR), total energy consumption, response time and resource utilization. The proposed uRank‐TOPSIS and HSRLBA achieved an energy consumption of 325 KWh, response time of 15.42 s, Makespan of the 1150 s, TCR of 98% and resource utilization of 92%. The proposed framework maximizes TCR and resource utilization and minimizes the Makespan, energy consumption, and response time.

Hybrid Learning based Radio Resource Management in 5G Heterogeneous Networks

Ultradensification using different types of small cells (SCs) is one of the key enabling solutions to meet the multiple stringent requirements of 5G cellular networks. However, radio resource management (RRM) in ultra-dense heterogeneous networks (HetNets) is not easy due to interferences in multi-tiered architecture and dynamic network conditions. Interferences in 5G HetNets can be efficiently managed only through the techniques which are adaptive and self-organizing to handle dynamic conditions in 5G HetNets. In this article, a machine learning (ML) based self-adaptive resource allocation scheme is proposed based on the combination of independent and cooperative learning and evaluated for ultra-dense 5G HetNets. The proposed scheme aims to improve the QoS of all users associated with different network tiers in ultra-dense HetNets simultaneously. The proposed solution adaptively optimizes the SCs transmit power either through independent learning or cooperative learning based on the varying density of small cells to minimize the interferences and ensure minimum QoS requirements for all users in different network tiers. The proposed scheme not only maintains the minimum required capacities for QoS provision to all users simultaneously but has also shown a significant improvement in the capacities of users in different network tiers in high interference scenarios as compared to the use of a single learning scheme.

Improved Grant-Free Access for URLLC via Multi-Tier-Driven Computing: Network-Load Learning, Prediction, and Resource Allocation

Grant-Free (GF) access has been recognized as a promising candidate for Ultra-Reliable and Low-Latency Communications (URLLC). However, even with GF access, URLLC still may not effectively gain high reliability and millimeter-level latency, simultaneously. This is because the network load is typically time-varying and not known to the base station (BS), and thus, the resource allocated for GF access cannot well adapt to variations of the network load, resulting in low resource utilization efficiency under light network load and leading to severe collisions under heavy network load. To tackle this problem, we propose a multi-tier-driven computing framework and the associated algorithms for URLLC to support users with different QoS requirements. Especially, we concentrate on <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$K $ </tex-math></inline-formula> - repetition GF access in light of its simplicity and well-balanced performance for practical systems. In particular, our framework consists of three tiers of computation, namely network-load learning, network-load prediction, and adaptive resource allocation. In the first tier, the BS can learn the network-load information from the states (success, collision, and idle) of random-access resources in terms of resource blocks (RB) and time slots. In the second tier, the network-load variation is effectively predicted based on estimation results from the first tier. Finally, in the third tier, by deriving and weighing the failure probabilities of different groups of users, their QoS requirements, and the predicted network loads, the BS is able to dynamically allocate sufficient resources accommodating the varying network loads. Simulation results show that our proposed approach can estimate the network load more accurately compared with the baseline schemes. Moreover, with the assistance of network-load prediction, our adaptive resource allocation offers an effective way to enhance the QoS for different URLLC services, simultaneously.