Hello, we're studying cost models in big data and while doing an experimental measure of the io of workloads executed in hbase and workloads executed in Phoenix on tables stored in hbase we found that the IO is always equal to zero !!!
in our experiment, we are running hbase on file system (we re not using hdfs). When we observed the cost of those queries (in Phoenix) we got a variable result for the io. we found that the only dimension of the cost vector returned by Phoenix that changes is the IO dimension (CPU and memory are always equal to zero). we found that this result is unlogical and we wish to have a better understanding of the reason why we' re having those values. thank you for ur help, yours sincerely, Asma ZGOLLI PhD student in data engineering - computer science
