EarthCube Architecture

Title: EarthCube Architecture

Session Lead: Steve Richard (Arizona Geological Survey)

Speaker 1: Steve Richard

Presentation Title: EarthCube Architecture Forum Results

Notes:

• What is EarthCube?: a cyberinfrastructure that fosters new, transformational geoscience.  How? By enabling discovery, evaluation, and access.

• What is Architecture?: A high level view of a system that can be used as a guide to assess if something fits or does not fit.

• Architecture Needs To: provide framework, occupy middle ground between uncoordinated technology-bazaar model and the single monolithic system model, and describe the communications necessary and the objects that need to be exchanged.

• Guidelines: architecture dictates interfaces instead of technology; architecture should be decoupled from the components, which are modular.

Speaker 2: Ilya Zaslavsky (UCSD)

Presentation Title: Geoscience Enterprise Architecture for Research (GEAR)

Notes:

• The Science Enterprise: ask, collect, hypothesize, test, and document, curate, and disseminate.  Also increasingly, integrate, analyzes, and collaborate.

• Purpose and Scope of Conceptual Design:

  • Purposes: framework for different components to fall into place; a way to access gaps and priorities, and to bring different communities together on a common, easily understood platform.

  • Scope EarthCube: Defined through interactions with stakeholders.

• The pitfall: a quote from Paul Brubaker, and in essence, avoid investing resources to design something that is unusable.

• Inputs: white papers, domain workshop results, review of existing cyberinfrastructure, and surveys.

• Technical Approach: federation of systems approach; define high-level components and communication channels; and emphasis on communication, system analytics as feedback mechanism driving evolutionary development.

• Perspectives on EarthCube: open federation of systems, cross-domain information integration, efficient execution of research workflows, alignment of stakeholder interest, and scholarly communication.

• Communication, Metrics: System management depends on knowing what is working.  Use information is used to provide feedback.

• Content and Computation: Resource system, registry system, and data processing system.

• Information model and Design report outline are also presented.

• Concerns: top 5 are presented - hitting the right level of granularity, identifying necessary communication channels, balancing current and future requirements, constructing a self-organizing plug-and-play system, and inventorying and characterizing existing components.

• Implementation details will evolve with time, so different components will be coming together.

• EarthCube document repository: http://workspace.earthcube.org/document-repository

Speaker 3: Phil Yang (NSF Spatiotemporal Innovation Center, George Mason University)

Presentation Title: CD: Developing a Data-Oriented Human-Centric Enterprise Architecture for EarthCube

Notes:

• Project Objective: EarthCube as defined in 2012 is expected to act as research engine, resource management platform, computing service provider, and education platform.

  • Benefit: the design results will become a reference framework.

• Spiral approach with 4 “checkpoints”:

  • Spiral 1: 10% of all requirements.

  • Spiral 2: 50% of all requirements.

  • Spiral 3: 80% of all requirements.

  • Spiral 4: 100% of all requirements.

• Progress so far: Spiral 1 started in 2013, and there are activities in the first 3 spirals with most activities completed for Spiral 1.

• Science Driver: Initial design is based on user requirements of data, information, or computing resource.

• EarthCube EA Reference and Architecture document have been completed.

• Capabilities: 3 parts - end user, enabling, and resource (different categories of users contribute to the requirements of these 3 types of capabilities).

• Services: several services have been completed with many more on the way.

• End User Operations: Examples presented are resource discovery activity and governance activity.

• Next Steps: refine the ontology, dictionary structure and complete the volume at a certain level; focus on an example to demonstrate how to use the enterprise architecture design; integrate with EarthCube reference architecture; formalize the output and make it open access for the benefit of the Geoscience communities; continue to engage communities.

• Next EA Design workshop will be one of the days during ESIP Summer Meeting (July 14-17, 2015).

Speaker 4: Daniel Pilone

Presentation Title: EarthCube Conceptual Design, a SCalable commUnity Driven ARchitecture (SCUDAR)

Notes:

• Process: major steps are - identification, development, capture, and generation.

• Focused on the core issues that need to be addressed in terms of EarthCube Challenges & Needs: mainly, wildly diverse data; heterogeneity of data sources/format/distribution mechanisms/etc; interoperability and community use of data is critically important; and highly sophisticated desired use cases.

• The fundamental: Existing data needs to be searchable and discoverable in EarthCube.

• Data Provider, EarthCube CI, Data Analysis are the 3 major components of the architecture.

• “Interoperability is Key,” “Model Drive Information Architecture,” “Iteration and Evolution” are the additional design guidelines.

• Capturing and Communicating: TOGAF approach - processes; data/information model; application architecture, and technology architecture.

• Topics Discusses: Core, prioritized use cases; drivers and architectural principles; federated vs centralized architecture; governance models; mobility of computation and services; commitment to infrastructure management and sustainment; deployment model and support for integrating building blocks.

• More updates will be available in May.

Comments/Questions:

• Each of the architecture has its viewpoint, but these viewpoints do not necessarily match.  It might be good to review these viewpoints and their similarities/differences.

  • This review might be helpful in comparing the different architectural approaches.

• High level components for the framework might need to be defined/clarified.

• Review of the use cases/building blocks and the information from the reviews should be made available for other.

• How can the information regarding the architectural designs be made publicly available?

• What are the key ways to test the designs?

  • Groups responsible for use cases and testbeds should be collaborating, so that the designs could be tested accordingly.

• Radical thought: EarthCube already exists (based on the current systems); EarthCube evolution should be about the integration of these systems?

  • Related question: How to integrate data from different types?

• Even though that there are a lot of existing capabilities, they might not all meant or have the same aim to serve the purpose and scope of EarthCube.  As a result, it is still important to evaluate the architecture/capabilities/components, etc with EarthCube’s focus in mind.

• Policies/processes to produce a framework might be one of the most valuable outcome.