.net - TPL Dataflow pipeline design basics -
i try create well-designed tpl dataflow pipeline optimal using of system resources. project html parser adds parsed values sql server db. have methods of future pipeline, , question optimal way place them in dataflow blocks, , how blocks should use? of methods cpu-bound, , of them - i/o-bound(loading internet, sql server db queries). think placing each i/o operation in separate block right way on scheme:
what basic rules of designing pipelines in case?
one way choose how divide blocks decide parts want scale independently of others. starting point divide cpu-bound portions i/o-bound portions. i'd consider combining last 2 blocks, since both i/o-bound (presumably same database).
Comments
Post a Comment