ioNode middleware
Steve Lee
on Tom Hoffman's blog posting on the TinyZIS project
(http://tuttlesvc.teacherhosting.com/wordpress/?p=135).
Not only are the developers, www.phosphorix.co.uk, local to me, but I
realised I also had some brief contact with it as the SHELL project. I
thought I should summarise my understanding.
There's a summary at,
http://www.phosphorix.co.uk/media/press/ioNodeWhitePaper/,
and several papers are available on the SHELL project site,
http://www.educationaldevelopment.net/shellfinal/default.htm.
The JISC funded SHELL project aimed to facilitate transfer of student
records from FE colleges to a local University and synchronise
enrolments. It provides a centralised Learner Record database which is
available online to the learners (rather than on demand query and
assembly from multiple sources). This is a perhaps a subset of SIF
functionality, and while the project was specific in aim the ioNode
infrastructure was an attempt at a more generic solution by Phosphoric.
The (original) system design is hub based middle ware with agents much
like SIF. There is a single central database for all records with a web
UI for learners. It uses Web Service standards for discovery, connection
and communication, mainly SOAP and ebXML. The communications payload and
database use IMS schemas.
The source has subsequently been made available at www.ioagent.org, but
there does not appear to be an active FOSS project. It is implemented in
Java (Java is Community Source rather than FOSS).
My brief involvement was in interfacing the student records system (SRS)
to the agent. This was actually a server proving a simple file based
interface to us (bit like EDI). We had to generate the data from the
student record system and drop it. Several schema and operational
problems arose, for example we did not have a transactional access to
the SRS as implicit in the system design, only the current data set
state (using SQL to query).
The project run into problems, some technical, some management, some
resources and some political due to the many institutions having other
priorities and agendas and changing staff. Some success was achieved and
ioNode exists as is available as Open Source.
It is clear that interoperability depends on more than just technical
issues. Institutional business processes need to be aligned to some
extent as do business modelling such as the way enrolments are handled
(e.g. 1 year or all years). System flexibility is needed to handle
variations and map (RDF?). It is also critical to get common buy-in and
understanding whether a project or standard is being devleoped.
Steve Lee