Multi-Domain Network Slicing With Latency Equalization

Abstract

With network slicing, physical networks are partitioned into multiple virtual networks tailored to serve different types of service with their specific requirements. In order to optimize the utilization of network resources for delay-critical applications, we propose a new multi-domain network virtualization framework based on a novel multipath multihop delay model. This framework encompasses a novel hierarchical orchestration mechanism for mapping network slices onto physical resources and a mechanism for dynamic slice resizing. The main idea is to locally redefine the delay requirements on each network domain depending on the conditions in the rest of the network. Delays larger than threshold (debt) are allowed in certain domains if there is a possibility to compensate such excessive delays in other segments of the network that can transmit the messages with less latency (credit). This tradeoff or delay threshold redefinition on different segments of the route is referred to as network latency equalization. For performance comparison, minimum cost routing with latency constraints is used as a baseline. We show that our approach enables significantly better utilization of the network resources measured in the number of slices with the same latency requirements that can be accommodated in the network.

Multi-Agent Deep Reinforcement Learning for Cooperative Computing Offloading and Route Optimization in Multi Cloud-Edge Networks

“…Several studies have been conducted on security-aware task allocation under incomplete CSI [47]- [49]. In addition, several studies have been conducted on end-to-end network slicing or multi-domain network slicing to integrate the multiple controls [50]- [52]. Although much research has been done, optimizing task offloading in an endto-end network remains an unsolved problem and one of the major challenges.…”

Cloud and Edge Computation Offloading for Latency Limited Services

“…We assume that the transmission rate of the limited-latency slice relaying the task between the BS and MEC server is r s packets per second. The delay of sending a total of p n 1 packets over L M links between BS and MEC is d [28]. Similarly, relaying the computational result back to the UE needs d B,M n,2 = L M /r s + (p n 2 − 1)/r s , and thus the total delay between BS and MEC for relaying the task n is…”

Secure and Intelligent Service Function Chain for Sustainable Services in Healthcare Cyber Physical Systems

“…With respect to the last two sub-issues, they concentrate on deploying the requested network service, modeled by SFC, and fulfilling function requests and resource demands of the service. Consequently, multiple algorithms for SFC placement and scheduling have been proposed and developed [5][6][7][8][9][10][11][12][13][14][15][16][17][18], such as the greedy method based algorithm [7], Markov random walk model based algorithm [8], multi-attributes based algorithm [9] and so on. In addition, deep learning (DP) approach, originated from the known machine learning (ML), has been attempted to allocate computing and networking resources of virtual SFCs in recent years [13] [14].…”