Completion Report

Project Summary:

The purpose and deliverables of the project were to provide a set of environments and standards for use in further projects utilising Oracle SOA 11g. Systems will be ported to 11g as individual projects not as part of ITS095. The following is a list of the deliverables achieved by the project are as follows.

The main deliverables of the project were:

• Technical Architecture Design for 11g  platform
• SOA 11g Development Standards
• SOA 11g installations in the Dev, Test and Live environments
• Defined test suite to test basic functionality when deploying to a new SOA infrastructure

This was achieved by

• A review of Oracle SOA 11g to document development standards for future projects.
• Create Oracle SOA 11g DEV, TEST and LIVE environments for use in future projects to allow future development to progress with no prior environmental set-up required.
• A defined and documented build process for a SOA 11g installation.
• A re-usable Test suite was developed to check deployments
• An automated standard deployment and implementation mechanism was produced for use with all future SOA 11g projects.
• A standard security model was introduced for Webserver and application access.
• DR test failover operation was designed and tested successfully in conjunction with ITI using Veeam to facilitate the failover and failback.
• The project provided a generic Fault Handling Suite for 11g, an improved version the 10g version.
• The TAD and Operational documents were produced , reviewed and handed over to Production Management

Analysis of Resource Usage:

Staff Usage Estimate: 151 days

Staff Usage Actual: 334 days

Staff Usage Variance: 121%

Other Resource Estimate: 0 days

Other Resource Actual: 0 days

Other Resource Variance: 0%

Explanation for variance:

The main variance issue centred around the use of new technologies , the lack of knowledge of usage and implementation had a large effect..

• As staff were unfamiliar with the new technologies , much of the effort was directed  towards the deliverables resulted in a trial and error approach. This was especially true around server build and security implementations. 
• Information was not readily available from Oracle which required numerous iterations of research and trial and error.
• Responsiblity for the security work and implementations were not clarified prior to commencement of the work around security. This resulted in more time being used by the Development team than initially intended.

Key Learning Points:

• Although all technical staff involved in the project received formal training , this took place a long time before the project started to allow them to make full use of it. It would be advisable to align training and subsequent project work where the training will be applied, possibly planning in the training as part of the project.
• For Dev Tech, depending the size and nature of the project ,  it may be desirable to have initial small R&D tasks or a prototyping stage for the work to be carried out. This can feed into subsequent stages and estimates .This would be a  valuable thing to do for any project where we are bringing new technologies into the infrastructure. Doing this allows us to finalise or nearly finalise a design before implementing anything in dev test or live. This has been done successfully for Drupal, MyEd and Timetabling.
• Where the department is embarking on using new technologies, it would be advisable to assess whether external consultants could be useful at the project outset or, for example, at the end of prototyping stages, during the project. The Oracle consultants could have signed off on the early work in the dev environments which could have save a lot time later for the cost of some focused consultant time.. Potentially, a criteria could be developed as regards the perceived complexity of the project deliverables and the available experience and knowledge within the teams involved. For example, the recent visit by Oracle consultants to review the SOA 11g environments created by this project provided valuable recommendations for ongoing work.
• At pre-defined check points during the project, a deliverables v budget assessment could have been carried out, possibly using a recognised method such as MoSCoW.

Outstanding issues:

Two deliverables of the original scope were not delivered.

• Review performance monitoring tools.
• Review Load testing model required.

The COM007 project will be the first to use the new infrastucture and will carry out Load Testing and review monitoring on the system in question and review any SOA configurations or tuning as required.

Ongoing , new systems being placed on the infrastucture will assess their own requirements as far as load testing and tuning is concerned.