5G network slicing using SDN and NFV: A survey of taxonomy, architectures and future challenges

Computer Networks, Год журнала: 2019, Номер 167, С. 106984 - 106984

Опубликована: Ноя. 17, 2019

The increasing consumption of multimedia services and the demand high-quality from customers has triggered a fundamental change in how we administer networks terms abstraction, separation, mapping forwarding, control management aspects services. industry academia are embracing 5G as future network capable to support next generation vertical applications with different service requirements. To realize this vision network, physical be sliced into multiple isolated logical varying sizes structures which dedicated types based on their requirements characteristics (e.g., slice for massive IoT devices, smartphones or autonomous cars, etc.). Softwarization using Software-Defined Networking (SDN) Network Function Virtualization (NFV)in expected fill void programmable resources. In paper, provide comprehensive review updated solutions related slicing SDN NFV. Firstly, present quality business followed by description softwarization paradigms including essential concepts, history use cases. Secondly, tutorial technology enablers SDN, NFV, MEC, cloud/Fog computing, hypervisors, virtual machines & containers. Thidly, comprehensively survey industrial initiatives projects that pushing forward adoption NFV accelerating slicing. A comparison various architectural approaches practical implementations, adoptions deployment strategies is presented. Moreover, discussion open source orchestrators proof concepts representing contribution. work also investigates standardization efforts regarding softwarization. Additionally, article presents orchestration slices single domain across domains while supporting tenants. Furthermore, highlight challenges research directions networks.

Язык: Английский

---

Optimal VNF Placement via Deep Reinforcement Learning in SDN/NFV-Enabled Networks

IEEE Journal on Selected Areas in Communications, Год журнала: 2019, Номер 38(2), С. 263 - 278

Опубликована: Дек. 13, 2019

The emerging paradigm - Software-Defined Networking (SDN) and Network Function Virtualization (NFV) makes it feasible scalable to run Virtual Functions (VNFs) in commercial-off-the-shelf devices, which provides a variety of network services with reduced cost. Benefitting from centralized management, lots information about traffic resources can be collected SDN/NFV-enabled networks. Using powerful machine learning tools, algorithms designed customized way according the efficiently optimize performance. In this paper, we study VNF placement problem networks, is naturally formulated as Binary Integer Programming (BIP) problem. deep reinforcement learning, propose Double Deep Q Network-based Placement Algorithm (DDQN-VNFPA). Specifically, DDQN determines optimal solution prohibitively large space DDQN-VNFPA then places/releases Instances (VNFIs) following threshold-based policy. We evaluate trace-driven simulations on real-world topology. Evaluation results show that get improved performance terms reject number ratio Service Chain Requests (SFCRs), throughput, end-to-end delay, VNFI running time load balancing compared existing literatures.

Язык: Английский

---

A Survey of Intelligent Network Slicing Management for Industrial IoT: Integrated Approaches for Smart Transportation, Smart Energy, and Smart Factory

IEEE Communications Surveys & Tutorials, Год журнала: 2022, Номер 24(2), С. 1175 - 1211

Опубликована: Янв. 1, 2022

Network slicing has been widely agreed as a promising technique to accommodate diverse services for the Industrial Internet of Things (IIoT). Smart transportation, smart energy, and factory/manufacturing are three key form backbone IIoT. management is paramount importance in face IIoT with diversified requirements. It important have comprehensive survey on intelligent network provide guidance future research this field. In paper, we thorough investigation analysis its general use cases well specific including energy factory, highlight advantages drawbacks across many existing works/surveys current terms set criteria. addition, present an architecture focusing above services. For each service, detailed application requirements architecture, associated enabling technologies. Further, deep understanding orchestration AI-assisted operation, edge computing empowered slicing, reliability, security. presented identify corresponding challenges open issues that can guide research. To facilitate implementation, case study integrated factory. Some lessons learnt include: 1) it necessary explicitly service function chains (SFCs) applications along underlying VNFs/PNFs supporting such SFCs; 2) crucial guarantee both ultra-low latency extremely high reliability; 3) resource heterogeneous domains importance. We hope useful researchers engineers innovation deployment

Язык: Английский

---

The Road Towards 6G: A Comprehensive Survey

Опубликована: Янв. 1, 2021

<p>As of today, the fifth generation (5G) mobile communication system has been rolled out in many countries and number 5G subscribers already reaches a very large scale. It is time for academia industry to shift their attention towards next generation. At this crossroad, an overview current state art vision future communications are definitely interest. This article thus aims provide comprehensive survey draw picture sixth (6G) terms drivers, use cases, usage scenarios, requirements, key performance indicators (KPIs), enabling technologies. First, we attempt answer question “Is there any need 6G?” by shedding light on driving factors 6G, which predict explosive growth traffic until 2030, envision potential cases scenarios. Second, technical requirements 6G discussed compared with those respect set KPIs quantitative manner. Third, state-of-the-art research efforts activities from representative institutions summarized, tentative roadmap definition, specification, standardization, regulation projected. Then, identify dozen technologies introduce principles, advantages, challenges, open issues. Finally, conclusions drawn paint “What may look like?”. intended serve as enlightening guideline spur interests further investigations subsequent development systems.</p>

Язык: Английский

---

Deep Reinforcement Learning for Resource Management on Network Slicing: A Survey

Sensors, Год журнала: 2022, Номер 22(8), С. 3031 - 3031

Опубликована: Апрель 15, 2022

Network Slicing and Deep Reinforcement Learning (DRL) are vital enablers for achieving 5G 6G networks. A 5G/6G network can comprise various slices from unique or multiple tenants. providers need to perform intelligent efficient resource management offer that meet the quality of service experience requirements use cases. Resource is far being a straightforward task. This task demands complex dynamic mechanisms control admission allocate, schedule, orchestrate resources. Intelligent effective needs predict services' demand coming tenants (each tenant with slice requests) achieve autonomous behavior slices. paper identifies relevant phases in slicing analyzes approaches using reinforcement learning (RL) DRL algorithms realizing each phase autonomously. We analyze according optimization objective, focus (core, radio access, edge, end-to-end network), space states, actions, algorithms, structure deep neural networks, exploration-exploitation method, cases (or vertical applications). also provide research directions related RL/DRL-based management.

Язык: Английский

---

Recent Advances of Resource Allocation in Network Function Virtualization

IEEE Transactions on Parallel and Distributed Systems, Год журнала: 2020, Номер 32(2), С. 295 - 314

Опубликована: Авг. 17, 2020

Network Function Virtualization (NFV) has been emerging as an appealing solution that transforms complex network functions from dedicated hardware implementations to software instances running in a virtualized environment. Due the numerous advantages such flexibility, efficiency, scalability, short deployment cycles, and service upgrade, NFV widely recognized next-generation provisioning paradigm. In NFV, requested is implemented by sequence of Virtual Functions (VNF) can run on generic servers leveraging virtualization technology. These VNFs are pitched with predefined order through which data flows traverse, it also known Service Chaining (SFC). this article, we provide overview recent advances resource allocation NFV. We generalize analyze four representative problems, namely, (1) VNF Placement Traffic Routing problem, (2) (3) problem (4) Redeployment Consolidation problem. After that, study delay calculation models protection (availability) allocation, two important Quality (QoS) parameters. Subsequently, classify summarize work for solving generalized problems considering various QoS parameters (e.g., cost, delay, reliability, energy) different scenarios edge cloud, online provisioning, distributed provisioning). Finally, conclude our article discussion state-of-the-art topics related fields, highlight areas where expect high potential future research.

Язык: Английский

---

Online VNF Lifecycle Management in an MEC-Enabled 5G IoT Architecture

IEEE Internet of Things Journal, Год журнала: 2019, Номер 7(5), С. 4183 - 4194

Опубликована: Окт. 1, 2019

The upcoming fifth generation (5G) mobile communications urge software-defined networks (SDNs) and network function virtualization (NFV) to join forces with the multiaccess edge computing (MEC) cause. Thus, reduced latency increased capacity at of can be achieved, satisfy requirements Internet Things (IoT) ecosystem. If not properly orchestrated, flexibility virtual functions (VNFs) incorporation, in terms deployment lifecycle management, may cause serious issues NFV scheme. As service level agreements (SLAs) 5G applications compete an environment traffic variations VNF placement options diverse or networking resources, online approach is needed. In this article, we discuss management challenges that arise from such heterogeneous architecture, onboarding scheduling. particular, enhance intelligence orchestrator (NFVO) by providing: 1) a latency-based embedding mechanism, where VNFs are initially allocated appropriate tier 2) scheduling algorithm, instantiated, scaled, migrated, destroyed based on actual traffic. Finally, design implement MEC-enabled platform evaluate our proposed mechanisms real-life scenarios. experimental results demonstrate scheme maximizes number served users system taking advantage allocation core without violating application SLAs.

Язык: Английский

---

Reducing Latency in Virtual Machines: Enabling Tactile Internet for Human-Machine Co-Working

IEEE Journal on Selected Areas in Communications, Год журнала: 2019, Номер 37(5), С. 1098 - 1116

Опубликована: Март 21, 2019

Software-defined networking (SDN) and network function virtualization (NFV) processed in multi-access edge computing (MEC) cloud systems have been proposed as critical paradigms for achieving the low latency requirements of tactile Internet. While virtual functions (VNFs) allow greater flexibility compared to hardware-based solutions, VNF abstraction also introduces additional packet processing delays. In this paper, we investigate practical feasibility NFV with respect Internet requirements. We develop, implement, evaluate Chain-based Low ImplemeNtation (CALVIN), a low-latency management framework distributed Service Function Chains (SFCs). CALVIN classifies VNFs into elementary, basic, advanced VNFs; moreover, implements elementary basic kernel space, while are implemented user space. Throughout, employs mapping one per Virtual Machine (VM) MEC system. Furthermore, avoids metadata structure batch packets conventional Linux stack so achieve short per-packet latencies. Our rigorous measurements on off-the-shelf hardware demonstrate that achieves round-trip times from ingress point via two forwarding (one space space) server egress order 0.32 ms. indicate coding encryption feasible small 256 byte an budget 0.35 ms; whereas, large 1400 can complete coding, but not within

Язык: Английский

---

Adaptable and Data-Driven Softwarized Networks: Review, Opportunities, and Challenges

Proceedings of the IEEE, Год журнала: 2019, Номер 107(4), С. 711 - 731

Опубликована: Фев. 26, 2019

Communication networks are the key enabling technology for our digital society. In order to sustain their critical services in future, communication need flexibly accommodate new requirements and changing contexts due emerging diverse applications. contrast traditional networking technologies, software-oriented concepts, such as software-defined (SDN) network function virtualization (NFV), provide ample opportunities highly flexible operations, fast simple adaptation of resources flows. This paper identifies challenges adaptable softwarized introduces a conceptual framework adaptations networks. We first explain how contribute adaptability through functional primitives observation, composition, control. review wide range options fine-granular observations well composition control provided by SDN NFV. The multitude "tuning knobs" complicates decision making, which is main focus this paper. propose enhance with data-driven e.g., machine learning modules, resulting deep making modules can learn react changes environment, flow demands, so support meaningful Finally, we make case employing concept empowerment realize truly "self-driving"

Язык: Английский

---

Integration of D2D, Network Slicing, and MEC in 5G Cellular Networks: Survey and Challenges

IEEE Access, Год журнала: 2021, Номер 9, С. 37590 - 37612

Опубликована: Янв. 1, 2021

With the tremendous demand for connectivity anywhere and anytime, existing network architectures should be modified. To cope with challenges that arise due to increasing flood of devices/users a diverse range application requirements, new technologies concepts must integrated enable their benefits. Service providers business companies are looking areas research enhance overall system performance. This article gives detailed survey about recent 5G technologies, solutions they provide, effect caused by addition current cellular networks. It is based on three most important concepts: Device (D2D), Network Slicing (NS), Mobile Edge Computing (MEC). study proposes design future networks integration all technologies. believed spectrum efficiency, energy throughput will greatly improved using D2D. The delay computational load reduced as tasks handled edge routers located at base stations. Thus offloading core capital expenses operational significantly slicing network.

Язык: Английский

---