Chapter 5. Harvesting

This chapter describes the typical steps that should be fulfilled when defining the Bootstrap Strategy and the subsequent actions required to successfully execute the Harvester.

The chapter walks through the process of:

  • Cataloging and classifying the existing Assets.
  • Importing Solution Packs.
  • Creating new Categorizations and Asset Types.
  • Configuring and executing the Harvester.
  • Submitting new Assets using the Asset Editor.
  • Searching Assets.

Use case

When Oracle Enterprise Repository (OER) is being implemented for the first time, one of the first questions asked by an organization is how can they bootstrap hundreds of existing Assets in a clean, organized, and exploitable way. There is little point in implementing OER, if hundreds (or even thousands) of Assets have to be manually entered into the tool. An even worse case scenario occurs when Assets are automatically loaded (harvested) but uncategorized, meaning that it is impossible to make sense of the information. That said, OER relies heavily on the Harvester and the Solution Packs to perform such tasks.

Tip

Bootstrap or bootstrapping is referred to as the process of Harvesting Assets into OER for the very first time.

As described in Chapter 3, Introduction To Oracle Enterprise Repository, Harvesting is the process of gathering information from different design-time and runtime sources, and loading these into OER. Harvesters are tools that enable the OER to discover and register Assets from design-time sources such as JDeveloper, for example, or runtime repositories such as an SOA production instance.

Harvesting is a continuous process and can be done in a variety of ways. For example, by the command line, using ANT, WSLT, JDeveloper plugins, and even custom scripts employing the REX API.

Use case

Tip

Section 6: Configuring and Using Automated Harvesting in Design-time and Runtime Environments of the OER Configuration Guide elaborates exhaustively on the Harvesting features and techniques. Please refer to this guide for more advance details on the use of the Harvester: http://docs.oracle.com/cd/E23943_01/admin.1111/e16580/harvest.htm#CEGHBGIJ

Before bootstrapping the repository, it is imperative to have a good understanding of the as-is Asset state and document the same. By having an accurate view of what the Assets are, where they reside, and what their current state is, it will be possible to define the best Bootstrapping Strategy and thereby to employ the right Harvester and Harvesting technique.

In the case of Weir & Bell Telecom, the main Assets and Asset Types were well known and this can be seen in the Asset Relationship View available in Chapter 4, Initial Configuration. Furthermore, during the maturity assessment exercise of the Project Enablement phase, information about Assets such as type, location, and status was gathered, quantified, qualified, and mapped against the corresponding OER Asset Type, if available. This information was later used to define the right Harvesting and Bootstrapping Strategy. The following table provides a summary of the information gathered:

Asset Name

Document Type

Category

OER Asset Type

Amount

Location(s)

Qualification

SOA Strategy and Roadmap

Word Document

Governance Framework à

Strategy

N/A

1

Content management system (accessible via browser)

Up-to-date

Service Catalogue

Excel Spreadsheet

N/A

N/A = Strategy

1

Outdated. Unreliable.

SOA Reference Architecture

Word Document

Governance Framework à

Reference Architecture

N/A = Governance Document

3

Up-to-date.

Development Standards

Word Document

Governance Frameworkà

Standards

N/A = Governance Document

1

Up-to-date.

Design Standards

Word Document

N/A = Governance Document

1

Up-to-date.

Programming Standards

Word Document

N/A = Governance Document

1

Up-to-date.

Runtime Framework Detail Designs

Word Document

Governance Frameworkà

Framework Component Detail Design

N/A = Design Document

4

Subversion

Up-to-date.

System Use Case

Word Document

Project Deliverables à

Requirements

N/A = Requirement Document

20~

Subversion
  • Mix (some outdated, some up-to-date).
  • Several versions are available.
  • In many locations.
  • In SVN project /tags folder latest versions (which correspond to Production Version).

Functional Design

Word Document

N/A =Design Document

20~

SOA High Level Designs

Word Document

Project Deliverables à

High Level Design

N/A = Design Document

5~

Service Detail Design

Word Document

Project Deliverables à

Detail Design

N/A = Design Document

60~

Data Mapping Document

Spreadsheet

Project Deliverables à

Detail Design

N/A = Design Document

60~

Unit Test Plans

Word Document

Project Deliverables à

Test Plan

N/A = Design Document

60~

JDeveloper SOA 11g Application & Projects

Folders and XML Documents

Project Deliverables à

Code

N/A

150+

Eclipse OSB 11g Projects

Folders and XML Documents

N/A

150+

WSDLs

XML Documents

Artifact : WSDL

Unknown (Many)

Subversion

MDS

SOA Suite

OSB
  • Several versions available.
  • In many locations.
  • Latest versions in the production system (SOA Suite and OSB).

Schemas (XSD)

Artifact : XSD

XSLTs

Artifact : XSLT

SCA Composites

Composite

Fault Policies

Fault Policy

Other XML Documents

Adapters, Mplans, Rules, JCA, and others.

Many available in Solution Pack

Having elaborated an as-is view of the Assets state and their status, it was possible to elicit the most reliable Harvesting Strategy.

Tip

Information about which Asset Types and Repositories are supported by OER is available in the following sections of the OER Configuration Guide:

6.1.3 Artifacts/Products Version Matrix

http://docs.oracle.com/cd/E23943_01/admin.1111/e16580/harvest.htm#BABGCDID

6.2.1 Obtaining the Harvester

http://docs.oracle.com/cd/E23943_01/admin.1111/e16580/harvest.htm#CIHDJGJF

A.8 Asset Type Inventory

http://docs.oracle.com/cd/E23943_01/admin.1111/e16580/appx.htm#CEGEDEJA

Additionally you may refer to the OER Asset Type Catalog for a full list of the Asset Types available in the solution and base packs: http://www.oracle.com/technetwork/middleware/repository/overview/asset.xls

Based on the qualification made, the Harvesting Strategy was as follows:

  1. Bootstrap Assets by Harvesting the Oracle SOA Suite 11g, Oracle Metadata Services Repository (MDS), and Oracle Service Bus (OSB) using the command line utility. The decision was made because it was generally agreed that the production systems contained the latest working versions of all SOA Services implemented deployed to it.

    Alternatives such as Harvesting from Subversion or Artifact Stores were considered, but ultimately discarded as these repositories contained several versions of the same Assets in different folders, for example, trunk, tags, and branches, and it was concluded that identifying the latest version for each Asset would have been more difficult than getting the latest version from production.

  2. As documents such as Governance Standards, Framework Components, High Level Designs, Detail Designs, and Test Plans (among others) were not available in the production systems, these were added manually using the OER console. It would have been possible to develop a custom script using the REX API. However, given the limited number of this type of Asset, the cost involved in developing the script would have been greater than the cost of manual entry (in the end it only took around three days to add these manually into OER). Furthermore, during the project stages, business analysts, architects, and designers could continue to use the OER console to register Project Assets such as Use Cases, Functional Specifications, Design Specifications, and Test Plans.
  3. Once the Repository was fully bootstrapped, designers and developers would use the JDeveloper OER plugin to consume and register Assets during the development stages of the lifecycle.

    Tip

    When it comes to defining the right Harvesting/Bootstrapping Strategy one size does not fit all. Therefore, it is highly recommended that proper analysis is done and a decision is based on a qualification similar to the one described earlier.

Use case

Further to the Strategy, it was identified that the following Categorizations and Asset Types were not available in the Base Pack or Solution Pack and therefore had to be created in OER:

Categorizations:

  • Governance Framework.
    • Strategy.
    • Reference Architecture.
    • Standards.
    • Framework Components Detail Design.
  • Project Deliverables.
    • Requirements.
    • High Level Design.
    • Detail Design.
    • Test Plan.
    • Code.

      Asset Types:

  • Requirement Document.
  • Governance Document.
  • Design Document.

    Tip

    Most OER Asset Types allow for Asset-related Documentation to be added as metadata, but Weir & Bell Telecom preferred to define explicit Asset Types for key Document/Electronic Assets, and then define the relationship of these Document Assets to the actual Code. This would allow relationships to be graphically visualized using the Navigator applet (for example, a composite application relationship to its corresponding detail design, and so on) instead of these being hidden as a content of the tab.

The following sections will provide a comprehensive guide on how the Repository was Harvested and configured based upon Weir & Telecom's defined Harvesting Strategy.

..................Content has been hidden....................

You can't read the all page of ebook, please click here login for view all page.
Reset