Flexible Hybrid Cloud Infrastructure for Seamless Management of HuBMAP Resources

点击翻译按钮获取中文摘要

To provide for the Infrastructure Component (IC) of the HuBMAP HIVE, Pittsburgh Supercomputing Center (PSC) and the University of Pittsburgh (Pitt) will create a flexible and scalable hybrid cloud infrastructure in close coordination with the HIVE Collaboratory, the HuBMAP Consortium, and stakeholders world wide. This proposed IC will seamlessly connect HuBMAP with storage and co-located compute resources for depositing, accessing, querying, searching, analyzing, and visualizing both raw and processed HuBMAP data. Data and tools incorporated into the IC will be findable, accessible, interoperable, reusable (FAIR), and put together in reproducible, shareable workflows that can be run locally and on all major public clouds. The IC services will be flexible and extensible to incorporate new data, tools, technologies, security needs, federated identity and user roles via modular architecture of APIs and containers. These features enable tools to be both rapidly deployed and modifiable, defined as distinct functions with defined inputs and outputs, enabling solidly constructed repeatable pipelines. The IC will provide the open data platform for study of tissue organization and function through 3D biomolecular tissue maps in the spatial framework of the human body envisioned by HuBMAP. Our vision for the HuBMAP infrastructure is flexible, efficient delivery on HuBMAP’s goals, on sustainable platforms, by engaged and experienced leadership. We see the infrastructure component in a key service role to other HIVE components, TMCs, NIH program, and interacting with other atlas projects. Our design is based upon existing substantial persistent infrastructure supported by many projects with a 30-year history. At PSC, HuBMAP will be sustained for decades at reasonable cost and/or integrate with other programs. In addition to leadership and grounded technology expertise, we are proposing full-time project management and imaging and data science expertise within our team. These capabilities will allow the infrastructure component staff to engage as “ever-present” participants in any aspects of the discussions of the integrated collaboratory conceived by the HuBMAP program (e.g. HIVE, TMCs, NIH staff). The structure of the HIVE enables the infrastructure component to focus on delivery and the engagement required for delivery.