Concept

Load balancing
   Estimate distribution strategy – using machine learning
Data flow directivity – using machine learning
   high/low/same layer
Compensation of missing( or out of order) node data
   Prediction by using near nodes – using machine learning
Announce breakdown node (to main and nearby node)
   Re-path plan

Machine Learning

Chronical, spatial learning
   Compensation of data lose caused by malfunction of sensors
Prediction of downtime
   Defining relationship among the processes
Recognizing characteristics and predicting occurring time of processes in trouble
(SML, small/simple machine learning) module in each node
   Known factor :factors directly given by near nodes
   unknown factor - indirect: the result by factors affect to near nodes
mass(control whole system) ML(LML, large machine learning )
   Interaction between main LML and SML