Data Space DRAFT NSF Site Visit Data Space
- Slides: 8
Data. Space DRAFT NSF Site Visit Data. Space Review: Microsoft Presentation 8 Feb 2010 (v 5) 1
Why is Microsoft Interested in Data. Space • Large-scale data-driven sciences requires significant investments in hardware and software infrastructure • Advances in data storage, processing and information extraction are critical success factors, and fundamentally change how science gets done • Microsoft is interested in participating in developing breakthrough technologies • Microsoft is building a life-sciences platform (Amalga Life Sciences) with similar requirements • Including research on federated cross domain authentication and access control 2
Microsoft Research Experience • Individuals involved in this project have broad and relevant research experience. Participants’ include: – Dr. Steven White: 3 D registration and imaging platform architecture – Dr. Allen Brown: Life-sciences computational platforms and data security – Dr. Antonio Criminisi: Anatomical labeling from 3 D data and image recognition – Dr. Martin Will: Life-sciences data management, analytics, geometry and graphics processing. 3
Collaboration with Neuroscience Team • Microsoft will work with the Data. Space neuroscience team (e. g. , Dr Gabrieli’s group): – Applying labeling and registration on neuroimaging data – Enabling cross-site data sharing and reuse. – Testing and evaluating platform architectures for remote data analysis 4
Collaboration on Architecture • The Data. Space system will be developed in an open collaboration between teams to enable sharing of both data and algorithms while allowing parties to protect the privacy and integrity of each. • Microsoft’s Amalga platform is a federated and loosely coupled architecture ideally suited for research programs like Data. Space • Microsoft will encourage and support adoption of open standards (e. g. RDF, SPARQL) to facilitate interoperation of systems and sharing of data and algorithms. 5
Backup slide 6
Microsoft Approach to Claims Based Security Single sign on Decouple services from authentication Token: A cryptographically signed data unit Support different types of credentials that expresses one Support federated scenarios for cross or more claims. domain authentication and access control Secure Token Identity Service • Facilitate identity delegation s Provider (STS) m lc ai • Facilitate claim mapping with d en t ica te , A cq ui re ue iss 3) Au th 4) n ke o T To k en • • Service User 5) Invoke Service using the claims 1) Acquire claim requirements 2) Return claim requirements Service (Relying Party 7
Image /Ontology Mapping Scenarios Clinical (Amalga) Pharma (ALS) Neuroscience (Data. Space) Indexing/Search X X X QC/Validation X Cross-image (longitudinal) X Cross-image (population) X Visualization X Planning (e. g. surgery) X Data aggregation X X X Pathology detection X X X 8
