Understanding Data Intensive Systems Using Dynamic Analysis and Visualization

1
Understanding Data Intensive SystemsUsing
Dynamic Analysis and Visualization

• Nesrine NOUGHI

2
Motivation and Context

• Understanding up to 50 of the cost of the
  software maintenance
• Data intensive systems
• Intensive usage of data ( towards Big Data)
• More and more dynamic systems (ORM, Web
  Application, etc.)
• Consequences
• The database often occupies a central place
• Systems are more and more complex to understand

3
Objective
Understanding the communication between the
database and the programs becomes a prerequisite
to data-intensive system comprehension

• Focus Capture SQL queries at run-time and
  visualize dynamically their impact on the
  database to ease program comprehension

4
Approach and Methodology
Logical Visualization
Conceptual Visualization
Workflow Detection
5
Approach and Methodology

• Step 1 Logical Visualization
• Uses the logical schema as representation model
• Takes one SQL trace and logical schema as input
• Allows visualization at a low-level of
  abstraction (tables, columns, etc)

6
Approach and Methodology

• Step 1 Logical Visualization
• Example

Logical schema
Logical visualization
SQL trace
7
Approach and Methodology

• Step 2 Conceptual Visualization
• Uses the conceptual schema as representation
  model
• Takes one SQL trace, both conceptual and logical
  schemas as input
• Allows visualization at a higher-level of
  abstraction (concepts, relationships)

8
Approach and Methodology
Step 2 Conceptual Visualization Example
Logical schema
Conceptual schema
Mapping
Conceptual visualization
SQL trace
9
Approach and Methodology

• Step 3 Workflow Detection
• Takes several SQL traces as input
• Allows the visualization of the data manipulation
  workflow of the program

10
Ongoing work

• Implementation of the first two steps

Logical visualization
Conceptual visualization
Conceptual visualization
Logical visualization
Conceptual visualization
loop detection algorithm
11
Future directions

• Complete and enrich the conceptual visualization
• Define an experimental protocol for the
  evaluation of the prototype with real users
• Use this visualization in order to
• animate the workflow of the program
• provide program documentation and diagrams