Research On Protocol Adaptation And QoS Guarantee For Multi-Mode Electric Internet Of Things

As an important space carrier for green development and low-carbon transformation,smart parks have become an important part of the construction of new power systems under "dual carbon" background.Multi-mode electric internet of things(EIoT)can provide communication and digital support for the low-carbon operation of smart parks and provide technical support for the construction of new power systems through deep integration of power and communication network.The"dual carbon" response of smart park is characterized by diverse communication protocols,and there is heterogeneity in various communication protocols,while the current mainstream EIoT cloud platform adopts message queuing telemetry transport(MQTT)protocol for data transmission.Therefore,it is necessary to research the multi-mode EIoT protocol adaptation and realize unified cloud platform data uploading and forwarding.Quality of service(QoS)guarantee is of vital importance to improve power communication service level in smart parks.MQTT provides three QoS levels.It is necessary to dynamically select MQTT QoS levels to satisfy the differentiated QoS requirements of multi-mode EIoT services.However,protocol adaptation and QoS guarantee in multi-mode EIoT still faces several challenges.First,multi-mode EIoT heterogeneous protocol adaptation and interoperability between traditional and constrained networks remain to be researched.Second,lack of closed-form models of delay and packet-loss ratio for three MQTT-specific QoS levels.Finally,EIoT services have differentiated QoS requirements.How to satisfy differentiated QoS requirements for low-carbon services under incomplete information is a pressing issue.In this paper,a communication architecture of multi-mode EIoT is firstly established.Based on publish/subscribe mechanism,the sender and receiver based on EIoT heterogeneous protocols are decoupled to realize MQTT-based unified cloud platform data uploading and forwarding.Further,caching mechanisms and load merging strategies for protocol conversion in cloud platform agents are investigated to achieve efficient interconnection between traditional and constrained networks.Second,the closed-form models of delay and packet-loss ratio for three MQTT-specific QoS levels are derived.Then,based on the upper confidence bound(UCB)algorithm in reinforcement learning,a delay-reliability-aware MQTT QoS level selection(DR-MQLS)algorithm is proposed.The QoS level selection problem is modeled as a multi-armed bandit(MAB)problem,the three QoS levels are abstracted as arms,so that the gateway can realize the dynamic MQTT QoS level selection based on local information such as delay,packet-loss ratio,whose optimization objective is to minimize the weighted sum of packet-loss ratio and delay.Finally,the simulation results validate that,compared with single and fixed QoS level selection strategies,DR-MQLS performs superior in packet-loss ratio and delay.