In recent years, with an increasing number of Internet of Things (IoT) devices, general cloud computing mode is hard to process large amounts of data with high Quality of Service (QoS). Edge computing is put forward to relieve the pressure of cloud servers, but most of them only focused on allocating tasks depending on cloud servers or edge servers with the virtualization technology. Resource-constrained smart mobile terminals (RC-SMTs) produce most of the data to be processed but some of them are usually not able to support even Docker technology. The cooperative computation capacity of RC-SMTs is potential but is often neglected by most researchers. However, there is little research focus on edge computing only among RC-SMTs without computing ability supported by servers. For this reason, this article proposes a framework named data-drive task offloading with a unified resource model (DDTO-URM) to manage the limited resource of IoT which enables the allocation of tasks constantly generated from the edge of the network. Then, a meta-heuristic algorithm called grouped crossover genetic algorithm (GCGA) is designed to obtain task offloading strategy under a resource-constrained environment. As a result, the computation capacity of the system is enhanced to cover the requirement by improving the utilization of RC-SMTs. Through the analysis of simulation, the proposed approach can deal with the problem of DDTO-URM better than benchmark algorithms under constraints, ensuring the real time and ultralightweight of the collaborative edge-computing system.

The expansion of Internet of Things (IoT) spawns large on-device machine learning demands, while the machine learning can be a hard task for resource constrained IoT terminals with fragmented data set. Federal learning (FL), which aims to build a joint model across multiple devices, IoT-FL has now become a promising path for learning on terminals. In broad FL fields, current server–client pattern cannot jump out of the third-party self-trustless problem, and recent researches suggest that even sharing training results may also reveal the raw data sets. Homomorphic encryption (HE) is a powerful method in privacy preserving, while so far it is hard to apply HE into multiparty computing (MPC) scenarios including FL. Combining with the existing IoT architecture, in this article, we customize a scheme (FL chain) dedicated for the privacy and trustiness issues in IoT-FL scenarios, which integrates blockchain smart contract and HE. Differ from traditional schemes, our FL chain is highly adaptive with current IoT architecture and it is the first scheme that applied HE into IoT-FL privacy preserving. Theoretical analysis and experimental results prove the feasibility of FL chain.