Euclid's US Science Data Center : Lessons Learned from Building a Small Part of a Big System

摘要

Euclid is an ESA M-class mission to study the geometry and nature of the dark universe, slated for launch in mid-2022. NASA is participating in the mission through the contribution of the near-infrared detectors and associated electronics, the nomination of scientists for membership in the Euclid Consortium, and by establishing the Euclid NASA Science Center at IPAC (ENSCI) to support the US community. As part of ENSCI's work, we will participate in the Euclid Science Ground Segment (SGS) and build and operate the US Science Data Center (SDC-US), which will be a node in the distributed data processing system for the mission. SDC-US is one of 10 data centers, and will contribute about 5% of the computing and data storage for the distributed system. We discuss lessons learned in developing a node in a distributed system. For example, there is a significant advantage to SDC-US development in sharing of knowledge, problem solving, and resource burden with other parts of the system. On the other hand, fitting into a system that is distributed geographically and relies on diverse computing environments results in added complexity in constructing SDC-US.