LitE: Load Balanced Virtual Data Center Embedding for Energy Efficiency in Data Centers

Abstract

Network virtualization (NV) enables efficient management of physical network (PN) resources by partitioning them into virtual data center requests (VDCRs), consisting of interconnected virtual machines (VMs) and virtual links (VLs). A key challenge in NV is virtual data center embedding (VDCE), which allocates PN resources to VMs and VLs and is -hard problem. Existing VDCE strategies often fail to balance energy efficiency and resource distribution, leading to sub-optimal solutions with higher energy consumption in data centers (DCs). This work presents LitE, a load-balanced VDCE strategy focused on minimizing energy consumption in single-domain PN. LitE uses a resource management strategy that considers server utilization, overloading probability, and energy consumption to select suitable servers for VM embedding. It then applies Dijkstra's shortest path algorithm for VL embedding to optimize energy use. Experiments show LitE improves energy efficiency by compared to baseline methods through better resource utilization. © 2025 Copyright held by the owner/author(s).