Reengineering Systems for MultiChannel Access:

1
Reengineering Systems for Multi-Channel Access

• A Systematic Literature Review

By Clive Jefferies, supervised by Pearl Brereton
2
Agenda

• Background
• Multi-Channel Access
• Replace or reengineer?
• Types of reengineering
• Systematic Literature Review
• Review planning
• Reporting
• Next steps

Image from Google images.
3
Background

• There are a large number of existing or legacy
  systems within companies and institutions.
• A lot of these systems are designed with a single
  channel of delivery in mind e.g. desktop
  application or web application.
• What if they wish to offer these systems as
  mobile solutions offering their functionality
  over many channels?

4
Background gt Multi-Channel Access

• Multi-Channel Access (MCA) - A software service
  that can be accessed by heterogeneous devices
  e.g. mobile phone, PC, PDA etc.
• Use case Medical professionals could query a
  patient data system at the point of patient
  contact.

Image from Google images.
5
Background gt Replace or reengineer?

• Service enablement is considered less risky than
  replacing an entire system e.g. undocumented
  business processes
• Service enablement also requires less resources
  e.g. time and people.
• Single to Multi-Channel reengineering problem

?
6
Background gt Types of reengineering

• Black box
• No knowledge of the code is needed
• Difficult when business logic is tied up with
  presentation logic
• More likely to be used as interim solution
• White box
• Source code is changed
• Takes longer than black box as knowledge of code
  needed
• Regarded as a more permanent solution

7
Systematic Literature Review (SLR)

• A protocol based literature review for gathering
  literature, extracting data and synthesising data
  in a systematic way.
• Thorough and less biased therefore increasing its
  scientific value (Kitchenham, 2004).
• 3 Stages Planning, conducting and reporting.

8
SLR gt Planning

• Identification of a need for the review
• Identify research questions
• Have there been any proposed frameworks for
  reengineering a system as a service?
• Are there any Service-Oriented reengineering
  methods or solutions that focus specifically on
  Multi-Channel Access?
• What are the issues that need to be addressed
  when attempting to reengineer a system as a
  service?
• Define protocol for conducting the review
• Gathering papers (search strategy, sources)
• Study selection (inclusion/ exclusion criteria
  and quality assessment)
• Extracting data from papers
• Synthesising data from papers

9
SLR gt Reporting

• 92 papers were found using the ACM, Keele
  Library, Google and IEEExplore.
• 28 of these were used for data extraction after
  the inclusion/ exclusion criteria were applied.

10
SLR gt Reporting gt Question 1

• Have there been any proposed frameworks for
  reengineering a system as a service?
• 15 Methods were found which had varying coverage
  of the lifecycle, ordering of steps and level of
  detail.
• Coverage examples
• SOR (Zhang et al., 2006) offers full coverage
• SMART (Lewis et al., 2005) focuses on analysis
  and design
• Ordering examples
• SoftLink (Sneed, 2003) starts with the source
  code
• SOAD (Zimmermann et al. 2003) businesses process
  required and source code compared

11
SLR gt Reporting gt Question 2

• Are there any Service-oriented reengineering
  methods or solutions that focus specifically on
  Multi-Channel Access?
• Two methods were found
• Reengineering toward a channel agnostic,
  conversational system, (Zimmermann et al. 2004)
• Multi-Channel Adaptive Information Systems
  (MAIS), (Comerio et al., 2004).
• 7 architectures/ solutions were found which will
  be analysed further.

12
SLR gt Reporting gt Question 3

• What are the issues that need to be addressed
  when attempting to reengineer a system as a
  service?

• Reengineering for MCA related
• Poor documentation will make the engineering
  effort difficult
• Granularity few system interactions for MCA
• Non-functional requirements (e.g. performance)
  mobile constraints
• Performance of Web services heavyweight
  protocols
• Security mobile users outside firewall

13
Next steps

• Need to compare Service-Oriented reengineering
  methodologies (ordering, level of detail,
  coverage ..)
• Need to evaluate MCA architectures to establish
  which best meets the constraints of mobile
  devices.
• XML and Web services are verbose. This could be a
  problem when they are used on mobile devices
  which have limited processing power,
  unpredictable connectivity etc. This needs to be
  investigated further.

14
References

• Kitchenham, B. (2004) Procedures for Performing
  Systematic Reviews. Joint Technical Report
  Software Engineering Group, Keele University,
  United Kingdom and Empirical Software
  Engineering, National ICT Australia Ltd,
  Australia.
• For methods references see abstract.