Applications. 1 Day s Offerings from CIO Online

Transcription

1 1 The Challenge of Quality in Multi-Tier, Multi-Language s Dr. Bill Curtis SVP & Chief Scientist CAST Software 1 Day s Offerings from CIO Online 2

2 2 Why Do New Features Take So Long? What results are you getting from your outsourcers? Source: Outsourcer Quality Survey Results Software Development Magazine March 2006 Better than in-house results Unusable or setback to progress Worse than in-house results 46% 14% 26% 8% Same as in-house results Developing the changes 3 19% 6% Documenting the changes Understanding the code 47% 28% Testing the changes How do maintenance staff spend their time? Source: Software Quality, Producing Practical and Consistent Software Mordechai Ben-Manachem Business Agility Is Limited by App Quality Ossified, contorted, complex legacy code Speed in responding to the market depends on the ease of modifying business applications 4 Well-engineered, high-quality code Harmfully complex Hard to change Easy to hack into Costly to maintain Secure and robust Easy to change Cheap to maintain Faster to build

3 3 Software Quality Iceberg TEST QUALITY EVALUATION degraded performance defects complexity security coupling cohesion architecture robustness QUALITY SYMPTOMS poor response time outages QUALITY CHARACTERISTICS interoperability excessive costs overruns program structure coding practices testability reusability maintainability scalability understandability flexibility 5 Steve McConell (1993), Code Complete. Godot s Gotta Be in There Somewhere Product Catalogue?th threads Resource pools? Connection pools? Error handling? Timeouts? Retail Website?K threads Order Entry? threads Credit Card? threads Shipping? threads Express Service? threads 6

4 4 Siloed Developer Skills Enterprise s Connectors Presentation Tier ERP Webmethods Web Services JSP / JSTL / JSF STRUTS - MVC Business Logic Tier Legacy s Batches COBOL CICS Connector Java Services Spring Data Tier EJB Hibernate - Ibatis Files Databases Batches C Pro C Database 7 Much of the complexity and risk of applications occurs at the interface between technologies beyond the comprehension of siloed skills Dispersed Development 8

5 5 The Devil Is in the Interface Enterprise s Middleware Web / Client Server s ASP/JSP/VB/.NET Presenta- tion Tier Web Services Legacy s CICS Monitor (Cobol) Tuxedo Monitor (C) CICS Connector Logic Java, C++, Frameworks Struts MVC, Spring Business Logic Tier COBOL Batch Shell Scripts Data Management Layer EJB Hibernate - Ibatis Data Tier Database Files Databases 9 Internal Quality Is Often Overlooked Quality The degree to which a product meets its specified requirements Problem Customers struggle to state functional requirements. They do not understand non-functional requirements. a failure to satisfy a non-functional requirement can be critical, even catastrophic non-functional t ti requirements are sometimes difficult to verify. We cannot write a test case to verify a system s reliability The ability to associate code to non-functional properties can be a powerful weapon in a software engineer s arsenal. 10

6 6 11 Stability Anti-Patterns and Patterns Attacks of self denial Bulkheads reduces impact Decoupling middleware Users exacerbates Chain reaction results from violating Steady state avoids Unbalanced capacities Fail fast Handshaking can avoid Blocked threads mutual aggravation Slow responses Unbounded result sets mitigates found near Integration points Cascading failures Timeouts Scaling effects SLA Inversion prevents Test harness Detects problems Circuit breaker works with Michael Nygard (2007). Release It! Quality vs. Code Quality Quality quality also measures how well the individual components work together to make up the overall business system Code Quality Code quality is the measure of individual components for compliance with standards d and best practices in the context of a specific language 12 Good code quality Good application quality

7 7 Supplementing Mature Processes Quality Engineering supplements CMMI to unlock even more business value from applications CMMI focus process improvement Six Sigma AQE focus product improvement Design for 6σ INNOVATE QUANTIZE STANDARDIZE STABILIZE Goal-driven improvements Statistical quality mgt. Tailorable quality processes Project quality practices 13 The 4 th Wave in Software Engineering 4 Product What: When: Why: Architecture, Quality characteristics, Reuse 2005 Ensure software is constructed to standards that meet the lifetime demands placed on it 3 Process What: CMM, ITIL, PMBOK, Agile When: Why: Provide a more disciplined environment for professional work incorporating best practices 2 Methods What: Design methods, CASE tools When: Why: Give developers better tools and aids for constructing software systems 1 14 What: 3 rd & 4 th generation languages, structured programming When: Why: Give developers greater power for expressing their Languages programs

8 8 How Do We Get to Dependable Software National Research Council Software for Dependable Systems As higher levels of assurance are demanded testing cannot deliver the level of confidence required at a reasonable cost. The cost of preventing all failures will usually be prohibitively expensive, so a dependable system will not offer uniform levels of confidence across all functions. The correctness of the code is rarely the weakest link. 15 Jackson, D. (2009). Communications of the ACM, 52 (4) Presenting Dependability Cases Objective Provide direct evidence that a system satisfies its dependability requirements stated as about Dependability goal Dependability claims Dependability properties argues provides Dependability case Dependability evidence 16 Auditable Complete Sound comprises Tests Proofs Analyses

9 9 17 Structural Analysis of Software Quality Internal Quality App. Internal Quality Immediate Impact On-Going Impact Health Factors Performance Robustness Security Transferability Changeability Quality Indicators Complexity Architecture Programming Practices Naming Conventions Documentation Size Quality Metrics Class complexity (Inh. depth) Class complexity (Inh. width) Method complexity (Param.) Method complexity (control flow) SQL Complexity Distribution Artifacts with recursive calls File conformity Dead code Structuredness Controled data access Coupling distribution Modularity Empty code Encapsulation conformity Inheritance Package naming Class naming Interface naming Method naming Attribute naming Constant naming Package comment Class comment Method comment Package size Class size (methods) Class size (attributes) Interface size Method size Ov ver 800+ architectural & languag ge-specific code checks Allianz Austria s Maintainability Results Business Need: Reduce the application management costs of Allianz Austria s insurance management system 300 RFCs and 700,000 lines of code modified every year used to serve 10 million customers and process 8 million claims a year. Solution Enforce Allianz specific application quality rules to stabilize maintainability score Automatically assess and monitor the quality and maintainability of applications Provide standard enforcement reports to help AD teams remediate problems faster Integrate application quality processes into Allianz Quality Management System Benefits Maintainability stabilized despite a 40% increase in code volume over 4 years Improved delivery on new functionality by 230% while reducing application maintenance costs by 20% over 3 years 56% reduction in defects in four years / KLOC Defects % reduction in defects in 4 years

10 10 Using Measures to Control Quality Objective 19 Measures should be managed as distributions Beware arbitrary thresholds derive them statistically Using Measures Diagnostically Contro ol Flow Complexi ity Departures from statistically derived relationships among the attributes of a component indicate the possibility of code pathologies Size (LOC, FP, E, etc.) 20

11 11 21 Business Value & Quality Business Value Minimize business risks Minimize IT costs Maximize business agility Optimize work productivity Maximize customer loyalty Tactical Objectives Maximize standards compliance Improve software readability Reduce vendor lock-in Reduce cost of ownership Reduce injected defects Reduce application mistakes Reduce application rework Ease team handoffs Minimize liquidated damages Maximize application availability Minimize degraded service Reduce learning curves Reduce modification effort Accelerate new function delivery Maximize application scalability Maximize speed of response Maximize information retrieval Minimize unwanted breaches Maximize customer confidence Maximize information protection Health Factors TRANSFERABILITY allows new teams to quickly begin working with an application CHANGEABILITY makes an application easier and quicker to modify ROBUSTNESS improves application stability & reduces injecting new defects PERFORMANCE Reduces degraded response times and increases scalability SECURITY affects an application s ability to prevent unauthorized intrusions Consortium for IT Software Quality CISQ IT organizations, Outsourcers, Agencies, Experts IT & AD Executives Technical experts 22 Define industry issues Drive standards adoption Buildappraiser program Create quality standard Developer certification Integrate with standards