How to Build an Effective Data-Mesh
摘要
Abstract
We propose a multi-tier architecture for scalable, tightly integrated management of heterogeneous storage implementations. After formalizing the problem and requirements, we analyze enabling technologies that address current shortcomings and introduce an architectural template suitable for adapted deployments. We implement this template in the multi-tier data-mesh, which unifies cloud and grid computing paradigms to provide policy-driven data placement, efficient data movement, and strong coordination across globally distributed storage resources used in megascience experiments.