/
C103 IT ecosystem

C103 IT ecosystem

Oct 19, 2010

Facilitated by Justin Montgomery, University of Chicago

General

  • Model contains different elements

  • Relations between elements are dependency oriented (x-to-y-to-z)

  • Elements and relations comprise high-level view of infrastructure

  • System shows distinction between "notional" thing and the "physical" entities to help users understand how systems are tied together

  • Can search for dependancies on a given system for a greater understanding of the impact of system outage

System

  • Created using Java

    • UI is Java Swing

    • Plans for UI to be a Web app, eliminating the need for Java client

  • Components are flexible, not server-dependent

  • System creates Graph for visual representation of the ecosystem

    • Graph can be exported and pulled into other systems

  • Database can be exported to an OWL file for human readability

  • Clusters and VMs are automated so users don't need to manually keep up with them

    • Some things, however, need human intervention to make certain data is accurate and up to date

  • Relational model has not been used in ordered to allow for future flexibility and extensibility without the need for blowing up DB and starting from scratch 

Discussion and Questions

  • Relations become obvious as they are being created

  • A Java framework has been developed to better leverage the system

  • Can the maintenance workflow be expanded?

    • Specifically to include import/export of new components/systems

  • Has there been buy-in from groups for data entry?

    • Meetings are held with other groups and changes are implemented by core group

    • Change management process is still under development

    • One issue is that at a certain level, it is only people who know what information needs to be included

  • Wisconsin is using Service Desk for a similar undertaking, but it has inherent weaknesses

    • A goal would be to get this application integrated with essential parts of Service Desk

  • APIs will be/are being developed to make automation possible/easy

  • Change Management process is critical to establish feedback loop

  • The network layer, displaying "what depends on what" is a critical component

    • This is good for reporting on what actual dependencies are, giving the organization a clearer picture of outage implications

  • Web service to allow browsing from disparate systems into the IT Ecosystem via links