Please includes and Daniel

At present our company as a result of the need of business assurance, plans to deploy two hadoop cluster, one is data clustering, to hold all data and provide the impala queries, the other is a compute cluster, all program will calculate the content of the migrated to the cluster, the deployment of the spark,

Use is spark runtime batch this batch need data was obtained from the data in the cluster, and then compute cluster operation to be performed, computing results to compute cluster hbase
Can you tell me whether the method is feasible, and whether there is no one has adopted the deployment plan
Still hope glad!!!!!!