Why pull-based pipelines are faster (by acco) — discussion

#databases #scaling