Back to the Future: Quo Vadis ICIS? Richard Bruskiewich
Back to the Future • Pragmatic concerns for development • Discussion… • Technology opportunity for I*IS • Discussion… • Quo Vadis ICIS? • Discussion… (*) insert acronym for your favorite crop
ICIS 5.* versus ICIS 6.0 • Two ICIS schemata are now in serious production use • Conflicting needs to support DLL and Windows apps versus new MBP • Duplication of effort dilutes resources and confounds coordinated development • Strategy for synchronization is needed
ICIS 6.0 • Is this major revision extensive enough? • ICIS 6.0 GEMS and OMS embrace the spirit of the GCP domain model • Should the rest of the ICIS schemata follow suit? • Should the table and field naming schemes of ICIS tables be reviewed and rationalized further? • “Best Practices” for table and field names(?)
The ICIS Product Bazaar • ICIS has a bazaar of software tools implemented in diverse languages: Fortran, C, Delphi, Java, etc. • Such tools are only interoperable under Windows via the DLL (probably now also somewhat available under Intel Mac OSX) • The GCP MBP is hosting “next generation” crop information system development with Java • But… ICIS standalones are currently delivering good value to the community, and Eclipse/RCP has not yet delivered
Catch-22 • MBP team slow to deliver promised workbench due to lack of dedicated software engineering staff on effort • Dedicated staff not available, due to commitment to a high level of maintenance engineering and new feature development on legacy ICIS Windows applications
The Strategic Question • Do the Windows-based ICIS standalone applications have functionality which cannot be (well) transferred to Eclipse/RCP? • e.g. integration with MS Excel(?) or equivalent • Do the “legacy” applications need to be abandoned, or just adapted?
Is there another way? • Maybe it is more important right now to standardize on the ICIS schema (for the time being) • Agree on one common ICIS 6.0 schema • Perhaps “legacy” Windows applications can have their life cycle extended by web service (dot.Net) integration of the DLL into the GCP MBP platform • Not the Davenport opinion, but…
Motivation • Maintaining computer server infrastructure is generally complex • I*IS installations are non-trivial to robustly establish and maintain • Complex software • Significant customization needs • Constantly evolving – version updates • Significant end user support needs (*) Insert your favorite crop acronym (i.e. R, W, M, Ch, etc.) here
Service Oriented Architecture • Concept: host major public I*IS crop centrals professionally hosted on “the Cloud” • I*IS informatics infrastructure is outsourced to 3rd party “experts”, so the ICIS community can focus on functionality and content The Cloud
What is “the Cloud”? • Cloud computing is a way of computing, via the Internet, that broadly shares computer resources instead of using software or storage on a local PC…. • …Details are abstracted from the users who no longer have need of, expertise in, or control over the technology infrastructure "in the cloud" that supports them… provision of dynamically scalable and often virtualized resources as a service over the Internet. http://en.wikipedia.org/wiki/Cloud_computing accessed March 4th, 2010
Hype Curve Dot-Com Gold Rush Cloud computing Dot-Com Bust (’90’s)
Benefits I The Cloud • Commodity IT infrastructure is outsourced to outside “experts”, so the customers can focus on functionality and content • 24 x 7, 99% uptime guaranteed. Server and network downtime is minimized • Customers (largely) don’t need to worry about resiliency of hardware and generic software (O/S) infrastructure – server and storage management (routine) backups - to support their applications
Benefits II The Cloud • One-stop application software release management • Can leverage generic cloud technology • e.g. Google Apps for • User authentication & authorization • Useful generic applications • Functional API for customer web applications • Web access to data is not prone to administrative headaches (one expert team manages all major installations securely and robustly on behalf of stakeholders)
Strategic Issues • How to maintain data privacy in between different customers of the I*IS service • Element of trust to be developed • Business continuity risk management • Technical challenges of moving ICIS onto the Cloud
What is ICIS? • What is ICIS now? • A collaborative community (“club”) developing an open crop information system? • A consensus crop data model & semantics? • e.g. Precise concept of germplasm identification • Shared sets of software? • Shared protocols and libraries • Database integrated suite of tools • All of the above? Perhaps… • But… what should it be in the future?
ICIS Strengths • Global public good • Recognized global expertise in crop information semantics • Achievements in solving important problems in crop data management • Germplasm tracking and naming (GMS) • Flexible data semantics (DMS) • Other useful tools: GEMS, LDMS, IMS, SMTA management, etc.
ICIS Weaknesses • Boutique informatics project • Coherence of system architecture is still suboptimal • Product bazaar becoming harder and harder to manage, extend and maintain • Fortran, C, Delphi, Java, etc… • Sometimes poorly coordinated with serious duplication of effort. Priority setting and community communication is relatively inefficient and often incomplete • ICIS community is more of a club than a robust professional business model • Human resource base for development is vulnerable to many counter-productive institutional cultural and fiscal variables
ICIS Opportunities • Reputation of ICIS in crop domain is still attracting significant stakeholder interest • Traditional ICIS community partners still around • Wider adoption within CGIAR (finally!) • Global Crop Diversity Trust (& other partners) • Continued evolution of crop science offers exciting new areas of growth that depend on traditional (crop pedigree) strengths of ICIS • Molecular breeding • GCP platform enablement empowers ICIS data integration with other important data sources
ICIS Threats • Loss of “traditional” partners • Bayer Crop Sciences (incl. Nunhems) lost interest • AAFC SPARC co-PI’s may soon retire? • Relevant well-funded initiatives may bypass ICIS • iPlant Collaborative? • Dissension exists within ICIS development team • Lack of implementation coherence and robustness • Dilution of resources • Lack of coherent and sustainable technical long term vision (other than within GCP/ MBP)
Prognosis • The evolutionary tree of software is littered with extinct and sub-optimal projects • ICIS may ultimately go the way of other pioneering and technically wonderful, but unsustainable (“extinct”) technical initiatives • To survive, prosper and better serve its client audiences, perhaps the ICIS community needs to move away from “ICIS Club” and towards “ICIS Business”