Potential-Game-Based 5G RAN Slice Planning for GBR Services

Multi-Operator Core Network (MOCN) is a radio access network (RAN) sharing scheme among multiple mobile network operators (MNOs), featured by the need that precious radio resources are shared by multiple dedicated core networks to achieve capital and operational cost reductions. The paper presents a study on a software-defined scheme of MOCN (SD-MOCN) to flexibly configure and control the radio resource allocation, in units of resource block group (RBG), based on a service level agreement (SLA) of RAN sharing among dedicated MNOs. The proposed SD-MOCN has been successfully realized with a RAN emulator, connected with and shared by two physical LTE core networks (CNs) with unique public land mobile network (PLMN) identities. Such a successful realization is showcased with both the functional and performance tests. Functionally speaking, cellular phones of each dedicated CN can successfully be connected with their associated PLMN core network via one RAN slice. Performance-wise, each MNO can reach its slice-based throughput and goodput linearly proportional to the number of allocated RBGs, considering the performance reduction factors caused by the effects of both the modulation/coding induced spectral efficiency and transmission overheads of user/control planes. The preliminary results of performance tests contribute three key effects: (1) slice isolation, (2) slice performance linearity, and (3) slice reduction factor linearity, where effect (1) can guarantee management independence of multiple dedicated CNs sharing a single RAN, effect (2) can ensure the linear response of a given RAN slice with respect to its allocated RBGs, according to the SLA planning for its dedicated CN, and effect (3) indicates that the performance reduction factors of the system-level throughput and goodput are both linearly applicable to the prediction of the slice-level throughput and goodput.

Network slicing is a fundamental feature of 5G systems that facilitates the provision of particular system behaviours adapted to specific service/application domains on top of a common network infrastructure. A network slice is in general composed by a core network slice and a Radio Access Network (RAN) slice. The realization of RAN slices is particularly challenging because it requires configuring and operating traffic differentiation and protection mechanisms to simultaneously deliver multiple and diverse RAN behaviors over a given pool of radio resources. In this context, this paper proposes to characterize the behavior of a RAN slice through the specification of a set of control parameters that are used to dictate the operation of the packet scheduling function at Layer 2 and the radio admission control function at Layer 3. An evaluation of the suitability of these parameters for achieving efficient radio resource sharing and isolation between RAN slices is presented when configuring a network for supporting a slice with multiple enhanced Mobile BroadBand services and another slice for providing Mission Critical services. The analysis reveals the different impact of the Layer 3 and Layer 2 parameters for isolating services of different slices depending on whether they require guaranteed or non-guaranteed bit rates.

Video is expected to be the dominant application by traffic volume over mobile networks in the near future. Mobile network operators are deploying video optimization techniques to enhance the user experience and network utilization for video delivery. Video optimization is typically “out-of-network” and includes techniques such as transcoding, transrating, time shifting and pacing that are implemented outside of mobile radio access and core networks. However, such techniques cannot easily exploit information about real time cell congestion and radio conditions for the video terminals because it is difficult to obtain such information from the radio access network. We propose a novel “in-network” video optimization technique, named Adaptive Guaranteed Bit Rate (AGBR), for HTTP-based Adaptive Streaming (HAS) video. This optimization technique is implemented at the base station and can thus exploit knowledge of the radio and congestion conditions. With only limited knowledge of the video stream properties or content, AGBR works by adjusting the throughput delivered to the different HAS clients that in turn adjust the video quality they request. The optimization algorithm maximizes aggregate quality across multiple video flows served by the base station without starving data clients, thereby improving the overall quality of experience. We demonstrate through extensive analytical modeling and simulations that AGBR can adapt to changing network conditions to support more video sessions at an acceptable quality than alternative algorithms, while enforcing fairness among all users competing for resources within a sector.

The virtualisation of Radio Access Networks (RANs) is one of the goals in designing 5G mobile networks. This paper aims at presenting a proof of concept for the virtualisation of radio resources using Open Air Interface (OAI), a software-based Long-Term Evolution (LTE) eNodeB physical emulator. OAI was extended to support multi-tenancy, representing diverse Virtual mobile Network Operators (VNOs) with different Service Level Agreements (SLAs). A comprehensive analytical model for managing the virtual radio resources has been proposed, with two key parts: estimation of available radio resources and their allocation to different VNOs. The estimation is performed by the model, and the allocation is managed by OAI scheduling. Various scenarios and use cases are studied in this virtual RAN environment, network performance being evaluated for different situations, by varying guaranteed levels, serving weights, and used services. Results show that the proposed approach offers almost the same capacity to guaranteed VNOs regardless of other existing VNOs, experiencing at worst a degradation of 32% of its initial allocated data rate, without violation of the guaranteed data rate. The data rate allocated to best effort VNOs may decrease up to 7% of its initial value, which is acceptable, to guarantee other more demanding SLAs.

In the context of 5G systems, the emergence of various use cases with diverse requirements has attracted great attention to network slicing. In a single physical network, several instances of logical end-to-end networks, i.e. slices, will be instantiated to fulfill these requirements. To this end, slices should share the resources of the physical network, which consist of Core Network (CN) and Radio Access Network (RAN) resources. Herein, we focus on the Radio Resource Management (RRM) in the context of network slicing. To maximize the pooling gains, dynamic resource sharing is preferred over static sharing. However, dynamic resource sharing can lead to undesirable inter-slice influences, in particular, the contention on radio resources. In this article, we show that, although the radio resources are dynamically shared among the users of different slices, proper slice management can realize slice protection. This is achieved by adjusting the fraction of radio resources allocated to the different slices by the Packet Scheduler (PS) and by limiting the number of users admitted to the network via Admission Control (AC). We propose an iterative algorithm to optimize the parameters of PS and AC in order to ensure that the service level agreements are satisfied. Extensive system-level simulations have shown that a central entity that tunes these control parameters can greatly increase the network's performance.

Network slicing is a key feature of forthcoming 5G systems to facilitate the partitioning of the network into multiple logical networks customised according to different operation and application needs. Network slicing allows the materialisation of multi-tenant networks, in which the same infrastructure is shared among multiple communication providers, each one using a different slice. The support of multi-tenancy through slicing in the Radio Access Network (RAN) is particularly challenging because it involves the configuration and operation of multiple and diverse RAN behaviour over a common pool of radio resources while guaranteeing a certain Quality of Service (QoS) and isolation to each of the slices. This paper presents a Markovian approach to model different QoS aware Admission Control (AC) policies in a multi-tenant scenario with Guaranteed Bit Rate (GBR) services. From the analytical model, different metrics are defined to later analyse the effect of AC mechanisms on the performance achieved in various scenarios. Results show the impact of priorities for services of different tenants and isolation between tenants when different AC polices are adopted.

5G radio access network (RAN) slicing is gaining momentum in finding applications in a wide range of domains, especially those requiring high-speed data transmissions such as space-air-ground integrated networks (i.e., drone-based systems). A key challenge in building a robust RAN slicing to support these applications is, therefore, designing a RAN slicing (RS)-configuration scheme that can utilize information such as resource availability in substrate networks as well as the interdependent relationships among slices to map (embed) virtual network functions (VNFs) onto live substrate nodes. With such motivation, we propose a machine-learning-powered RAN slicing scheme that aims to accommodate maximum numbers of slices (a set of connected VNFs) within a given request set. We present a deep reinforcement scheme that is called Deep Allocation Agent (DAA). In short, DAA utilizes an empirically designed deep neural network that observes the current states of the substrate network and the requested slices to schedule the slices of which VNFs are then mapped to substrate nodes using an optimization algorithm. DAA is trained towards the goal of maximizing the number of accommodated slices in the given set by using an explicitly designed reward function. Our experiment study shows that, on average, DAA is able to maintain a rate of successfully routed slices above 80% in a resource-limited substrate network, and about 60% in extreme conditions, i.e., the available resources are much less than the demands.

The 5th generation of mobile communications is expected to provide a network suitable for any service requirements using a technology termed network slicing. In network slicing, network resources are divided and are allocated to slices for services. Radio access network slicing is essential for end-to-end network slicing. The state of the radio access network changes from moment to moment and automatic control of network slicing is necessary to respond to service requirements in real time. Automatic control is applied using deep reinforcement learning that can learn optimal control by repeating trial and error. In this paper, we propose a method to allocate radio resources that satisfy the service requirements regardless of the number of slices using deep reinforcement learning. The proposed method flexibly controls multiple slices by observing the state of one slice and calling a model that controls one slice multiple times. The proposed method was evaluated by simulation using a scenario in which the number of slices changes with time. The results of the evaluation show that variations in slice numbers achieve almost completely satisfying of slice requirements without being affected by other slices.

Network slicing is one of the key enabling techniques for 5G, allowing Mobile Network Operators (MNOs) to support services with diverging requirements on top of their infrastructure. The MNOs should be able to offer network slices as a service and provide customisable and independent virtual networks to verticals. The slicing of an end-to-end (E2E) mobile network is divided into Core Network (CN) slicing, and Radio Access Network (RAN) slicing. In this paper, we assess the requirements for using radio hypervisors to enable RAN as a Service (RANaaS). We evaluate the current state-of-the-art on radio virtualisation with respect to these requirements and identify the missing features. Then, we present the eXtensible Virtualisation Layer (XVL), a software layer that provides the missing functionality for enabling RANaaS and can be added on top of existing radio hypervisors. We outline XVL's architecture and design choices, as well as evaluate its performance in terms of the delay to provision virtual radios, the delay introduced to forward IQ samples, and the computational overhead. Our results show that XVL enables leveraging existing radio hypervisors to support RANaaS.

Network Slicing (NS) represents a key technology enabler for advanced connectivity and data processing tailored to customers' specific requirements. While significant progress has already been achieved for Core NS, Radio Access Network (RAN) slicing still presents limitations in terms of sharing infrastructure, Service Level Agreement (SLA) guarantees, isolation, resource scheduling and allocation. In this context, this paper firstly introduces a novel slices configuration framework for the 5G New Radio (5G NR) infrastructure able to dynamically migrates the radio resources among the slices, while preserving the Quality of Service (QoS) of the served users. Our solution is illustrated in detail and tested on top of a real case 5G scenario, using a software-based simulator. Finally, this paper investigates the flexibility, scalability, and real-time properties of the proposed method, as required in the future 5G cloud-based architectures.

Network slicing is a promising technique for wireless service providers to support enhanced mobile broadband (eMBB) and ultra-reliable low-latency communication (URLLC) services in a shared radio access network (RAN) infrastructure. In this paper, we apply numerology, mini-slot based transmission, and punctured scheduling techniques to support eMBB and URLLC network slices. For efficient allocation of radio resources (e.g., physical resource blocks, transmit power) to the users, we formulate RAN slicing problem as a multi-timescale problem. To solve this problem and address the dynamics of the traffic, we propose a hierarchical deep learning framework. Specifically, in each long time slot, the service provider employs a deep reinforcement learning (DRL) algorithm to determine the slice configuration parameters. The eMBB and URLLC schedulers use their own attention-based deep neural network (DNN) algorithm to allocate radio resources to their corresponding users in each short and mini time slot, respectively. Simulation results show that the proposed framework can achieve a higher aggregate throughput and a higher service level agreement (SLA) satisfaction ratio compared to some other RAN slicing approaches, including the resource proportional placement algorithm, decomposition and relaxation based resource allocation algorithm, and distributed bandwidth optimization algorithm.

In 5G and beyond wireless systems, Network Slicing (NS) feature will enable the coexistence of extremely different services by splitting the physical infrastructure into several logical slices tailored for a specific tenant or application. In sliced Radio Access Networks (RANs), an optimal traffic sharing among cells is key to guarantee Service Level Agreement (SLA) compliance while minimizing operation costs. The configuration of network functions leading to that optimal point may depend on the slice, claiming for slice-aware traffic steering strategies. This work presents the first data-driven algorithm for slice-aware traffic steering by tuning handover margins (a.k.a. mobility load balancing) per slice and adjacency with a proportional controller based on SLA criteria. The tuning process is driven by a novel indicator, derived from connection traces, showing the imbalance of SLA compliance among neighbor cells per slice. Performance assessment is carried out with a system-level simulator implementing a realistic sliced RAN offering services with different throughput, latency and reliability requirements. Results show that the proposed algorithm improves the overall SLA compliance by 9% in only 15 minutes of network activity compared to the case of not steering traffic, outperforming two legacy mobility load balancing approaches not driven by SLA.

Multi-access Edge Computing fundamentals, services, enablers and challenges: A complete survey

To deliver on the vision of network slicing, future 5G networks will have to provide virtualization and isolation for a variety of Radio Access Network (RAN) technologies. In this paper we study how to implement RAN slicing in Wi-Fi networks. In particular, we present a scheduling algorithm that allocates airtime through a set of virtual Wi-Fi interfaces executing on the same or different physical radios, in order to fulfill each tenant's Service Level Agreement. We demonstrate a practical implementation of our scheduler on Linux systems based on the mac80211 driver that contrary to other existing solutions is independent from underlying hardware drivers. Finally, we experimentally evaluate the performance of our scheme through a set of realistic experiments.

End-to-End (E2E) network slicing enables wireless networks to provide diverse services on a common infrastructure. Each E2E slice, including resources of radio access network (RAN) and core network, is rented to mobile virtual network operators (MVNOs) to provide a specific service to end-users. RAN slicing, which is realized through wireless network virtualization, involves sharing the frequency spectrum and base station antennas in RAN. Similarly, in core slicing, which is achieved by network function virtualization, data center resources such as commodity servers and physical links are shared between users of different MVNOs. In this paper, we study E2E slicing with the aim of minimizing the total energy consumption. The stated optimization problem is non-convex that is solved by a sub-optimal algorithm proposed here. The simulation results show that our proposed joint power control, server and link allocation (JPSLA) algorithm achieves 30% improvement compared to the disjoint scheme, where RAN and core are sliced separately.