EarthCube - Brokering: A Way Forward for Global Multi-disciplinary Data Sharing

EarthCube – brokering – a way forward for global multi-disciplinary data sharing

Steven F. Browdy

http://earthcub.ning.com/groups/brokering

google doc session: http://goo.gl/vjl4Y

• Earthcube goal – to transform the conduct of research in geosciences by supporting community-based cyberinfrastructure to integrate data and information for knowledge management across the geosciences
  • If you built it – they will not come – so talk to the community
• Create a road map
  • How will work in the larger governance and other groups
• Expectation – want to engage the long tail folks
  • This is one of the primary aspects – all groups have this important
  • Long tail folks are the scientists on their own – outside of the larger data centers
  • Do pilot studies and hack-a-thons
    • Hack-a-thons are quick – 1 off experiments
• History of what done
  • Worked on roadmap – still evolving & align with other roadmaps
  • Had first hack-a-thon – used Euro GEOSS broker
  • Pilots – NSIDC arctic Pilot – ongoing – moving towards operation status
  • Attend June Charrette
  • Cross-fertilize with other earthcube groups
• Roadmap
  • Community
  • Information sciences and
  • Theorectical input
• Meta-summary
  • Infrastructure is a complex, dynamic process
  • Benefits – lower barriers, enhances multi-disciplinary  (requirement of earthcube)
• Community engagement
  • Use use cases to target science communities – need to do this in a flexible and evolutionary way
  • Look at small community problems
  • After cycling – look at larger problems – inter-community not just intra-community
• Hacka-a-Thon1
  • EuroGEOSS – info on ning site
  • Broker needs to be configured to access metadata and data
  • Provider of data had to provide the metadata and the data
  • Users did not have to register anything with the broker – just use tools
  • Talked about successes and failure during webex
  • Brokering is part of a robust, adaptive cyberinfrastructure
  • Next
    • Deal with governance
    • Continue and develop pilots and community engagement
  • People wanted broker to stay open – it has
  • Set up a wiki for discussion
  • Too narrowly focused – will include old hack-a-thon people
  • Extend length of hack-a-thon – but have meeting in through
• Pilot
  • NSIDC Arctic Polar Pilot
• There are lots of different ideas about what brokering is
  • Worked on common use cases
• Possible broker approaches
  • Single broker
  • Single broker in cloud
  • Multiple instances of same broker in cloud
  • Deploy in community – no communication between communities
• EuroGEOSS Broker
  • Deployed in many situations – GEOSS common infrastructure, WIS (WMO information system) access, and light semantic mediation
  • New accessor – interoperate with data providers
  • Uses ISO 19115 data mode
• Broker use cases
  • Tradition query, semantics-embed query, access on a common grid, third-party clients
• Brokering can include – publishing, discovery broker, MNL component, semantic broker
• Q – if you want to involve more sophisticated sementatic inside or add to broker
  • It would depend on the broker – can build it in, but from architecture – not sure why want to hide behind the broker
• Principals – whereas:
  • Diverse, use data differentially, different tools, access differently, providers are asked to provide data outside normal communities
  • Brokers should facilitate access to both big end and long tail
• We believe:
  • No one technology/standard will serve all user needs
  • Brokers can be effect at achieving flexibility
  • Brokering/framework – provides cyberinfrastructure
  • Maintained by brokering not by data users or provides
• Definitions of brokering framework
  • Brokers are middleware connect client and server in earthcube infrastructure
  • Facilitate run-time interconnection
  • Different components and functionality (ex. Data access, processing, publishing)
• Definition 2
  • Enables at runtime service components
• Q – how loosely or tightly coupled the components will they be
  • This is not defining that – semantic capability – it does not have to be there
  • Semantics should be outside the broker
  • In earthcube – is there a broker or a type of broker
  • In the future –have interoperable brokers that are used by earthcube
• It would be useful to access broker from different systems – what is the level of decoupling for different types of brokers – in semantic it would make sense to couple
  • Want decoupled – don’t know design yet
• Current USGS broker – it has light components – both internal and external components
• When you talk about interoperability – genetic – talk about data – but semantic is goal – is it important to define level of interoperability
  • There are lots of other types of interoperability, semantic interoperability is left to semantic group
• Q – everyone seems to be waiting on some other group
  • Needs to be community driven – so need to engage community to make progress
  • Governance team – will have something in place this year… will take time
• Q – is reference architecture – same as other groups showed
• Reference architecture
  • Domain infrasture – has service bus – includes metadata, polices, models
    • Within community – serve long tail
  • Cube – consists of multiple components
    • Broker, governance and stands, c
  • Q (Kelly) Should cross-domain registries be diff color – not as certain as other sections of infrastructure
  • Collaborative environments – social network
  • Brown cloud behind grey cloud
    • Brown cloud – high performance platforms and additional protocols
      • Will access/consume earthcube infrastructure (but issue of how infrastructure will work with high performance)
    • Grey cloud is web
• Governance questions
  • How should the brokers be deployed and maintained
  • Policies for brokering across boundaries
  • Published data, information, and knowledge standards – is there a convergent set of standards
  • What is process for evolving and innovating CI
• Use cases
  • Need to be interdisciplinary
  • Wants to use full range
  • Q – do we have common use cases for earthcubeu
    • Some groups have and some haven’t
    • Working on compiling the use cases
    • Need to be a primary activity to inform
• Future brokering possibilities
  • 2 main area – add functionality and different architecture/framework of how to deploy
  • Broker interoperability
    • May need new standards, or publishing capabilities, more governance issues
• Q – have you thought about existing standards for brokers ex. Wps and there are existing mappings on how to chain brokers
  • Existing standards/processes need to be use – but haven’t done this yet
• Q do you have a registry of brokers
  • It is not a large list – has been around for a long time – there are other areas other than science that have used them for a while – but not long list
• Q – can you list 3 brokers
• Q – is ECHO a broker
  • No
• Future – second and 3rd hack-a-thon mid-Aug and mid-Sept, more pilots
• ACTION – if have broker email [email protected]
• There is also a list serve www.arthcube.ning.com/groups/brokering