Real Proct Reliable Protocol Conformance Testing with Real

Real. Proct: Reliable Protocol Conformance Testing with Real Nodes for Wireless Sensor Networks Junjie Xiong 2011 -11 -15 1

Outline n n Motivation Real. Proct Design Evaluation Conclusion 2

Motivation n n Wireless sensor networks (WSNs) are vulnerable to software bugs. Simulation is different from real execution. Testbeds are designed for network performance evaluation rather than software bug detection. Large-scale real deployment is expensive. Real. Proct uses a small number of real sensor nodes to mimic large-scale WSNs and test the protocol against the specification. 3

Outline n n Motivation Real. Proct Design Evaluation Conclusion 4

Challenges n n Sensor nodes are resource-constrained devices. Sensor node is difficult to control than a computer. Volatile wireless environment in WSNs will result in random packet loss. How to test the protocol with various topologies and events when Real. Proct only employs a few real sensor nodes. 5

Protocol Conformance Testing Process n PCT process q IUT (Implementation Under Test) 6

Real. Proct Architecture Point of Control & Observation SUT (System Under Test) Upper Tester Lower Tester 7

Topology Virtualization 8

Event Virtualization n Packet disorder 9

Test Execution and Verdict n n Suppose packet loss probability is L 0, a test case is executed n times, and it passes n 1 times and fails n 2 times. If n 1 > n 2, declare as pass, the false negative probability is Let the FN error rate be lower than Th, then the minimum n is 10

Outline n n Motivation Real. Proct Design Evaluation Conclusion 11

Performance Evaluation n Two real Telos. B sensor nodes and a PC. Contiki 2. 4: μIP TCP/IP protocol. Two new bugs and previous bug repetition. q q q Bug 1 & 2 – Connect to opened & unopened TCP ports. Bug 3 – An SYN/ACK packet loss. Bug 4 – SYN packet duplication. 12

Performance Evaluation n Bug 1 – Connect to opened TCP ports. q Test opened port 0 & 80 (within 0 to 65535). 13

Performance Evaluation n Bug 1 – Connect to opened TCP ports. q Test opened port 0 & 80 (within 0 to 65535). 14

Performance Evaluation n Bug 2 – Connect to unopened TCP ports. 15

Outline n n Motivation Real. Proct Design Evaluation Conclusion 16

Conclusion n As a protocol testing tool, Real. Proct finds two new bugs, discovers all the previously detected bugs in the TCP/IP stack, and validates the Rime mesh routing protocol. n Propose two techniques, topology virtualization and event virtualization, for testing. n Design an algorithm to tackle the inaccuracy problem caused by non-deterministic events in test execution. 17

Q&A Thank you! 18