Introduction
In-network computing achieves higher attention in both the network and computing areas, where network devices are endowed with computing capability, compared to only being in charge of transferring data in the traditional form [1]. The primary force of this change is to integrate resources within the network to reduce the cost of task processing. Following this trend, the edge in-network computing paradigm has emerged and is increasingly adopted. In contrast to in-network computing, it specifically refers to resource integration at the edge, which means integration of the computing, caching, and communication resources between terminals and the remote cloud, providing services with less delay and energy consumption. This paradigm makes multi-domain heterogeneous resource management possible and provides a solution for edge nodes' cooperation. It is evolving with the development of 5G and 6G and has become a promising trend for decentralized latency-sensitive applications. However, as the edge nodes are geographically distributed and have limited capabilities, the collaborative management of edge-side resources mainly faces two challenges:
The trust assurance of multi-domain resource integration
The overall schedule of heterogeneous and geographically distributed edge resources