A Delay-Aware Approach for Distributed Embedding Towards Cross-Slice Communication

The evolution of modern applications and the adoption of the 5G network slicing architecture require minimizing the communication delays. Virtual Network Functions (VNFs) offer low latency services to remote users. In the context of Edge Computing, these virtualized services are hosted in Edge Clouds, which, however, have limited resources. Embedding the Network Slice requests from users is a challenging problem, and demands distributed solutions in the presence of resource and delay constraints in the Edge Network. This paper introduces a distributed delay-aware service discovery mechanism, which performs cache-based forwarding, trying to efficiently discover VNFs in the Edge Clouds (ECs) and enable Cross-Slice Communication (CSC); that is the sharing of common VNFs between different Network Slices (NSes). Furthermore, the distributed NS embedding (NSE) problem is addressed by a modified k-Shortest Path (k-SP) approach to minimize the total round-trip network delay under the capacity constraints of the underlying infrastructures. The proposed method significantly decreases the execution time by 70% to provide the distributed NSE solution compared to the baseline k-SP algorithm.