Toward Using Reinforcement Learning for Trigger Selection in Network Slice Mobility

Abstract

Recent 5G trials have demonstrated the usefulness of the Network Slicing concept that delivers customizable services to new and under-serviced industry sectors. However, user mobility's impact on the optimal resource allocation within and between slices deserves more attention. Slices and their dedicated resources should be offered where the services are to be consumed to minimize network latency and associated overheads and costs. Different mobility patterns lead to different resource re-allocation triggers, leading eventually to slice mobility when enough resources are to be migrated. The selection of the proper triggers for resource re-allocation and related slice mobility patterns is challenging due to triggers' multiplicity and overlapping nature. In this paper, we investigate the applicability of two Deep Reinforcement Learning based algorithms for allowing a fine-grained selection of mobility triggers that may instantiate slice and resource mobility actions. While the first proposed algorithm relies on a value-based learning method, the second one exploits a hybrid approach to optimize the action selection process. We present an enhanced ETSI Network Function Virtualization edge computing architecture that incorporates the studied mechanisms to implement service and slice migration. We evaluate the proposed methods' efficiency in a simulated environment and compare their performance in terms of training stability, learning time, and scalability. Finally, we identify and quantify the applicability aspects of the respective approaches.

Toward Network Slicing Enabled Edge Computing: A Cloud-Native Approach for Slice Mobility

“…Approach Proposed [13] Software-Defined Networks [6] Reinforcement Learning [1] Inter-Slice Control Function [3] Deep Reinforcement Learning [2] MEC Slicing Integration Framework [14] Deep Learning [15] Kubernetes (Velero) [16] Unified MEC Slicing Framework [17] Software-Defined Networks [18] Novel Centralized Control Plane Algorithm [19] Novel End-to-End Network Slicing Architectural Framework Our Proposed Software-Defined Networks proposals for new frameworks and architectures for network slicing and MEC integration, and slice resource management. However, network slice initiation and mobility challenges introduced by their integration in a highly dynamic and mobile environment demand further exploration.…”

“…However, the approaches do not consider mobility challenges. Addad et al in [6] proposed a reinforcement learning approach for the selection of optimal actions to refine the slice mobility decisions and network slice resources. Mlika et al in [3] proposed a deep reinforcement learning-based solution to deal with the resource allocation problem in MEC-enabled vehicular networks.…”

“…Though several emerging use cases and services for 5G and beyond networks rely on the integration between network slicing and MEC, the development of the technologies required for network slicing-enabled MEC is still in its early stages [1], [2]. Network slices are provisioned with dedicated resources to meet service-specific performance requirements; however, due to the limited computational capacity of the MEC servers, the dedicated network slices are not readily available on each MEC server [6]. Therefore, it is essential to develop effective resource allocation policies and new methods to deal with network slice mobility.…”

Online Service Migration in Mobile Edge with Incomplete System Information: A Deep Recurrent Actor-Critic Learning Approach

“…Wu et al [5] considered jointly optimizing the task offloading and service migration, and proposed a Qlearning based method combing the predicted user mobility. Addad [41] et al proposed an enhanced network function virtualization edge computing architecture that incorporates Q-learning based methods to implement service and slice migration. These works considered the case where the decision-making agent knows the complete system-level information.…”

Deterministic Service Function Chaining over Beyond 5G Edge Fabric

“…The authors in [10] studied the joint SFC deployment and resource management problem (JSDRM) in heterogeneous edge environments to minimize the total system latency and proposed a scheme based on a game model to jointly deploy SFCs and manage resources. The authors in [11] [12] also investigated the resource re-allocation of SFC strategy due to different mobility patterns. Different from the existing work, aiming at minimizing the service latency and network cost in the literature, deterministic latency assurance [2] [13] is more crucial for the packets of the time-critical flows.…”