Risks to Database Activities

Transcription

1 The Measured Performance of Department of Computer Science Allegheny College University of Pittsburgh, 2007 In Conjunction with Mary Lou Soffa (UVa/CS), Panos Chrysanthis (Pitt/CS), Bruce Childers (Pitt/CS)

2 Outline Introduction to Database Applications Motivation What is a Database Application? 1 Introduction to Database Applications Motivation What is a Database Application? 2 Traditional Software Testing A New Testing Paradigm 3 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 4 Experiment Design Instrumentation Costs Coverage Monitoring Costs

3 An Interesting Defect Report Motivation What is a Database Application? Database Server Crashes When you run a complex query against Microsoft SQL Server 2000, the SQL Server scheduler may stop responding. Additionally, you receive an error message that resembles the following: Date Time server Error: Severity: 1, State: 0 Date Time server Process 52:0 (94c)... An Input-Dependent Defect This problem occurs when one or more of the following conditions are true: The query contains a U N I O N clause or a U N I O N A L L clause that affects many columns. The query contains several J O I N statements. The query has a large estimated cost. BUG (SQL Server 8.0)

4 An Interesting Defect Report Motivation What is a Database Application? Database Server Crashes When you run a complex query against Microsoft SQL Server 2000, the SQL Server scheduler may stop responding. Additionally, you receive an error message that resembles the following: Date Time server Error: Severity: 1, State: 0 Date Time server Process 52:0 (94c)... An Input-Dependent Defect This problem occurs when one or more of the following conditions are true: The query contains a U N I O N clause or a U N I O N A L L clause that affects many columns. The query contains several J O I N statements. The query has a large estimated cost. BUG (SQL Server 8.0)

5 Real World Example Motivation What is a Database Application? A Severe Defect The Risks Digest, Volume 22, Issue 64, 2003 Jeppesen reports airspace boundary problems About 350 airspace boundaries contained in Jeppesen NavData are incorrect, the FAA has warned. The error occurred at Jeppesen after a software upgrade when information was pulled from a database containing 20,000 airspace boundaries worldwide for the March NavData update, which takes effect March 20. An Important Point Practically all use of databases occurs from within application programs [Silberschatz et al., 2006, pg. 311].

6 Real World Example Motivation What is a Database Application? A Severe Defect The Risks Digest, Volume 22, Issue 64, 2003 Jeppesen reports airspace boundary problems About 350 airspace boundaries contained in Jeppesen NavData are incorrect, the FAA has warned. The error occurred at Jeppesen after a software upgrade when information was pulled from a database containing 20,000 airspace boundaries worldwide for the March NavData update, which takes effect March 20. An Important Point Practically all use of databases occurs from within application programs [Silberschatz et al., 2006, pg. 311].

7 Motivation What is a Database Application? Program and Database Interactions P select update m insert delete Basic Operation Program P creates SQL statements in order to view and/or modify the state of the relational database D1 De

8 Database Interaction Granularity P Motivation What is a Database Application? R 1 A B C D m i D k R 2 E F G H m j D l I R 3 J K L Database Interactions Program P interacts with two relational databases D k and D l at different levels of granularity (relation, record, attribute,...)

9 Types of Applications Motivation What is a Database Application? Database Centric Applications Interaction Approach Program Location Embedded Interface Inside DBMS Outside DBMS Testing framework relevant to all types of applications Current tool support focuses on Interface-Outside applications Example: Java application that submits SQL Strings to an HSQLDB relational database using a JDBC driver

10 Outline Introduction to Database Applications Traditional Software Testing A New Testing Paradigm 1 Introduction to Database Applications Motivation What is a Database Application? 2 Traditional Software Testing A New Testing Paradigm 3 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 4 Experiment Design Instrumentation Costs Coverage Monitoring Costs

11 Traditional Software Testing A New Testing Paradigm Focus on Testing Individual Components P Input Output a 5 print... exit Final Result: 45 Byte Code Database Virtual Machine File System Graphical Interface Operating System Traditional Assumption Defects may exist in program P and/or P s execution environment

12 Traditional Software Testing A New Testing Paradigm Various Approaches to Software Testing Structural Testing Random Testing Input Space Software Reliability MTTF(P) = k Freq Input Specification Testing Techniques and Supporting Tools Structural testing requires a test coverage monitor!

13 Testing Environment Interactions Input P Traditional Software Testing A New Testing Paradigm Output a 5 print... exit Final Result: 45 Database Virtual Machine File System Operating System A New Direction in Software Testing Defects may exist in P s interaction with its environment. This suggests the need for a database-aware test coverage monitor!

14 Outline Introduction to Database Applications Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 1 Introduction to Database Applications Motivation What is a Database Application? 2 Traditional Software Testing A New Testing Paradigm 3 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 4 Experiment Design Instrumentation Costs Coverage Monitoring Costs

15 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Coverage Criteria for Database Applications P m i 3 delete from R where A > 100 define(r) 6 select * from R use(r) Candidates for Coverage Monitoring Find defects in the database interactions by ensuring that the test suite covers all of the possible def-use associations and/or calling contexts

16 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Challenges of Database-Aware Monitoring Id Test Case (testlistfiles) 1. Signature Method Under Test (listfiles) 2. SQL Select Path... /usr/bin/bibtex /usr/bin/bison Path(s) 3. Result Set Files Relation SQL Statement select Path from Files where ucase(path) like %/usr/bin/bi% Testing Challenges Traditional coverage monitoring does not reveal how the test case causes the method to interact with the database

17 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Overview of the Coverage Monitoring Process Test Suite Adequacy Criterion Program Instrumentation Database Instrumented Program Instrumented Test Suite Test Coverage Monitoring Test Requirements Coverage Results Adequacy Calculation Current Considerations Focus on the design, implementation, and performance evaluation of the instrumentation and coverage monitoring components Adequacy Measurements

18 Database-Aware Coverage Trees Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Test Case Method Invocation Database Interaction Point Database Relation Attribute Record Attribute Value Instrumentation Probes Use static and dynamic (load-time) instrumentation techniques to insert coverage monitoring probes Coverage Trees Store the coverage results in a tree in order to support the calculation of many types of coverage (e.g., data flow or call tree)

19 Comparing the Coverage Trees Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Tree Characteristics Tree DB? Context Probe Time Tree Space CCT Partial Low - Moderate Low DCT Full Low Moderate - High DI-CCT Partial Moderate Moderate DI-DCT Full Moderate High Table Legend Database? {, } Context {Partial, Full} Probe Time Overhead {Low, Moderate, High} Tree Space Overhead {Low, Moderate, High}

20 Database-Aware Instrumentation Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Test Coverage Monitoring Instrumentation Interaction Location Interaction Type Program Test Suite Defining Using Defining-Using Important Goal Efficiently monitor coverage of database state and structure without changing the behavior of the program under test

21 Phases of Coverage Monitoring Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes Coverage Tree Initialization Start Testing Before Update Probe Database Interaction After Probe End Testing Coverage Tree Storage Update Coverage Tree Coverage Tree Monitoring Operations Database-aware probes: Capture the SQL String Consult the database schema and result set meta-data Extract and analyze portions of the database state Update the coverage tree

22 Relational Differencing Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 2 rel j 3 A 2 2 rel j t 1 t 2 t 3 A Before t 1 t 2 t 3 A 1 A After Handling Database Modifications The probes use relational differencing to determine that record t 2 and attribute value t 2 [2] were modified by the SQL U P D A T E command

23 Outline Introduction to Database Applications Experiment Design Instrumentation Costs Coverage Monitoring Costs 1 Introduction to Database Applications Motivation What is a Database Application? 2 Traditional Software Testing A New Testing Paradigm 3 Coverage Monitoring Basics Fundamentals of Coverage Monitoring Instrumentation Probes 4 Experiment Design Instrumentation Costs Coverage Monitoring Costs

24 Experiment Design Instrumentation Costs Coverage Monitoring Costs Characterizing the Case Study Applications Test Suites Application # Tests Test NCSS / Total NCSS R M /548 = 50.5% F F /558 = 59.1% P I /579 = 35.1% S T /620 = 58.9% T M /748 = 47.5% G B /1455 = 52.8%

25 Experiment Design Instrumentation Costs Coverage Monitoring Costs Details about the Database Interactions Interaction Counts Application executeupdate executequery Total R M F F P I S T T M G B

26 Static Instrumentation Costs Experiment Design Instrumentation Costs Coverage Monitoring Costs FF PI RM ST TM GB All 10 Static Instrumentation Time sec FF PI RM ST TM GB All Application Attach probes to all of the applications in less than nine seconds Statically inserting probes increases space overhead

27 Experiment Design Instrumentation Costs Coverage Monitoring Costs Coverage Monitoring Time: Static Versus Dynamic Time Overhead Instr Tree TCM Time (sec) Per Incr (%) Static CCT Static DCT Dynamic CCT Dynamic DCT Discussion Static has poor space overhead but leads to a minimal increase in testing time. Static is less flexible than dynamic.

28 Experiment Design Instrumentation Costs Coverage Monitoring Costs Further Comparison of Static Versus Dynamic 14 Norm Sta CCT Sta DCT Dyn CCT Dyn DCT TCM Time sec Discussion Norm Sta CCT Sta DCT Dyn CCT Dyn DCT Instrumentation Technique Tree Type GB Static is faster than dynamic / CCT is faster than DCT

29 Experiment Design Instrumentation Costs Coverage Monitoring Costs Varying Database Interaction Granularity Discussion Time Overhead DB Level TCM Time (sec) Per Incr (%) Program Database Relation Attribute Record Attribute Value Static supports efficient monitoring since there is a 53% increase in testing time at the finest level of interaction

30 Conclusions and Future Work Experiment Design Instrumentation Costs Coverage Monitoring Costs Concluding Remarks A new perspective on software testing and an efficient and effective database-aware test coverage monitor Future Work Perform demand-driven instrumentation Use the coverage tree to reduce or prioritize a test suite Conduct experiments with larger database applications Resources