Computation Task Offloading Mechanism For IoT Applications In Smart Communities

In recent years,the Internet of things(IoT)has been booming,and the applications of IoT such as live streaming and augmented reality emerge endlessly.To make the living environment more "digital" and "intelligent"has become the goal of people.The smart community uses technologies such as the Internet and the Internet of Things to integrate the smart home system of the family with the IoT system and services of the community,forming a new mode of community management.There are many new IoT applications with high delay sensitivity in smart community,such as interactive games,smart home,smart health care and so on.To shorten the service response time and improve the "intelligent" experience is a goal of the development of smart community.At the same time,with the growing demand of society and families for energy,environmental pollution,energy depletion and other problems are becoming more and more serious.While committed to improving the " intelligent " experience,smart community also takes green and sustainable as another development goal.As a key technology to play the role of intelligent devices in the IoT,edge computing moves computing,storage and other resources from the cloud to a location closer to users,which solves the needs of IoT terminal devices for computing performance,storage resources,energy efficiency and other aspects.However,some of the existing general solutions are not specific enough for smart community scenarios.On the one hand,the existing computation offloading mechanism that focuses on latency does not consider the response time extension and resource waste caused by the repeated transmission of the same data file during offloading.The common way is to solve the computation offloading problem and the data cache problem separately,which ignores that the data file cache will affect the data transmission time and then affect the offloading decision.On the other hand,in the research on the use of green energy,the existing literature is more focused on the single mobile device with energy harvesting device,discussing the trade-off between quality of service and available power.These literatures rarely consider the fixed intelligent terminal devices(such as smart cameras,smart security devices,etc.)in the smart community,so the existing research is not fully applicable in this new scene.In view of the high latency sensitivity of some IoT applications in the smart community,this paper considers adding a sharable cache to the edge server to further reduce the latency caused by the data transmission process,and proposes a computation offloading mechanism based on the edge sharable cache.Firstly,a value-based edge cache update method(VECUM)is designed to maximize the average cache value and efficiently use the limited cache space.Then,a two-scene computation offloading method based on greedy strategy(TSOGS)is designed.The latency gain is set as greedy strategy,and the offloading decision making and computation resource allocation are studied based on the determined data file cache state.Finally,simulation results show the effectiveness of the proposed mechanism compared with the other four comparison algorithms.Aiming at the scenario of IoT applications based on green energy supply in smart community,this paper designs a more flexible computation offloading mechanism based on local energy consumption perception.Firstly,a dynamic adjustment method of local energy consumption unit cost is designed to encourage the use of green energy,control the consumption of non-green energy,and make the offloading decision process more flexible and environmentally friendly.Then,a computation offloading method based on game theory and Lagrangian multiplier method is designed.The decision of computation offloading is studied by constructing a multi-user game model,and the resource allocation is studied by optimizing the local computing cost and the edge computing cost respectively,and the offloading decision is updated iteratively based on the resource allocation results until Nash equilibrium is reached.Finally,simulation results prove the effectiveness of the proposed mechanism compared with the other three comparison algorithms.