EarthCube Conceptual Design - Scalable Community-Driven Architecture

EarthCube Conceptual Design - Scalable Community-Driven Architecture

Interactive session; goal: get feedback on concept -- are we heading the right direction?

Introduction

Emily Law, JPL

Setting the Context

• EarthCube: trying to construct a community-driven dynamic cyberinfrastructure that will support:

  • standards for interoperability

  • infuse advanced tech to improve and facilitate interdisciplinary research

  • enable geoscientists to address challenges by fostering a community-governed effort to develop a common cyberinfrastructure

  • collect, analyze data…

• funded projects include domain end-user workshops, conceptual designs, building blocks, research coordination networks (RCNs), integrative activities (IAs), governance

• there are a number of committees:

  • science committee, tech & architecture comm. (TAC), council of data facilities, liaison team, engagement team (& also leadership council)

  • many of these committees have working groups within them; within TAC there’s science use cases, gap analysis, standards, testbed, architecture

• architecture working group workshop in June

  • high level principles defines, white papers on architecture options for community and TAC review, architecture recommendations for EarthCube

Overview

• want to provide a high level system blueprint for definition, construction, and deployment of both existing and new components to ensure that they can be unified into structure for EarthCube

• methodology

  • using architecture development lifecycle methodology

    • ID stakeholders, their concerns/needs

    • ID use cases, main drivers

    • selection of architecture frameworks & development of arch. principles

    • development of arch. models

    • capture the arch. artifacts into consolidated report

    • generate rec’s for adoption of architecture for EarthCube program

• EarthCube can provide visualization platform to compare model & data and place to perform analysis

• guiding principles

  • scalability (horizontally, vertically), community-driven, open science, interoperability, sustainability, distributed, data model driven

• needs to support diverse sources of data

• needs to provide data analytics platform

• need to take a broader view of all activities in data lifecycle so can provide all of the right functionalities to support the EarthCube community (data generation → data curation → data transport → data ingest → data management → discovery, access, distribution → data analytics → visualization)

• common model, and then each domain has a domain-specific model on top of that

• Questions

  • data model driven architecture vs information model -- same? yes.

  • are funded building block projects part of this architecture? we have yet to do that, but it is in our plan

  • the scope of this is enormous -- are you actually _building_ this or just integrating or…? hopefully both; want this to guide implementation; want to support entire data lifecycle from end to end for a long time

  • this is a conceptual architecture/abstraction; this is not a referenced architecture where you specify standards, API...

Discussion

led by Dan Pilone, Element84 & Emily Law, JPL

• Richard Gabriel (Carnegie Mellon) study of ultra large scale systems → how we get from systems of systems to repository, etc.

• tension between community focus or “benevolent dictator” model for designing this system

• stakeholders

  • do we have the right ones? missing any key ones?

    • careful about all stakeholders being EarthCube people bc they’re funded to figure out what EarthCube is--could be like chasing your own tail

  • do they overlap at all? too much?

  • are they useful to provide use cases and personas that help drive the system?

  • a lot of interoperability problems are solved by people figuring out how to make things work for a concrete project and then you tap that resource to solve problems in other contexts--so would caution against being too precise at ontological level because don’t think that works

• What are the community priorities?

  • addressing interoperability question!

    • interoperability is not just about metadata, it’s about ability, software, implementation...

  • converting to knowledge has more to do with human activity than infrastructure activity

  • how to maximize compatibility between what’s going on in EarthCube community and repository community

• what provision would make it possible to incorporate social, economic, other non-geoscience data into this?

  • science data has specific attributes; other data have their own specific attributes, but that doesn’t mean that some of the attributes aren’t in common -- we’ll have to look at specific examples

• To wrap up, we are looking for input and these opinions are really helpful

• Where we are:

  • soliciting reviewers for conceptual architecture document (NOW!); frank opinions are welcome

  • then will incorporate feedback and review comments

  • write actionable recommendations and incorporate into final rec

• if you’re interested in participating, contact emily at [email protected]