Node Resource Management Model Of Hierarchical Edge Computing

Edge computing has become one of the key techniques in Internet of Things(Io T)solutions along with the rise of Io T.Hierarchical edge computing can be used in the Industrial Io T(IIo T).The paper focuses on hierarchical edge computing in IIo T which consists of static deployment and dynamic resource scheduling.Static deployment shows how to allocate hardware resources which is constrained by cost when hierarchical edge computing needs new deploying.Computing capacity allocation could be one of the problems when a new hierarchical edge computing network is deployed.A hierarchical edge computing capacity allocation method based on M/M/1c queueing model is proposed.Global average delay and loss are marked as performance indicators.The paper proposes Optimized Hierarchical Capacity Allocation Algorithm(OHCAA)to minimize global average delay when the system loss is lower than a certain threshold.The paper also makes some simulations to verify that OHCAA could decrease global average delay and loss.Dynamic resource scheduling requires the network to promote hardware utilization during runtime in order to get better performance.Deploying edge computing with existed devices has a problem that it is difficult to modify hardware.The paper models this problem with queueing theory and proposes Dynamic Link Model based on Nodes Relation(DLMNR)to achieve resource balance based on nodes relations.The paper also proposes an offloading strategy in the cache queue.Result analysis is given when delay performance which is related to environmental factors is analyzed.Meanwhile,the paper also makes simulations which are implemented by Java.Simulation results show that DLMNR has better global delay and its loss is acceptable.Finally,the paper prospects the research.In the meantime,the paper proposes some problems that were not solved yet.There are two aspects.First,tasks should be offloaded to proper nodes based on their features.Second,there may be compatibility issues between offloading and load balancing.The paper has 31 figures,9 tables and 57 references.