Can Software Implemented Fault-Injection be Used on Real-Time Systems?
Software Implemented Fault Injection (SWIFI) is a well-established technique for fault injection, but with a significant drawback for Real-Time Systems: intrusiveness, also known as “probe effect”. In fact, for most fault models, additional code has to be run on the same processor that executes the application. The danger lies in some deadlines being missed as a consequence of that overhead.
This paper identifies the sources of intrusiveness, and discusses the procedures to measure it. The question of what level of intrusiveness can be considered acceptable is also addressed.
A Pentium version of an existing SWIFI tool (Xception), developed with no real-time considerations in mind, was tested on a system composed by off-the-shelf (COTS) components (a standard PC with a Pentium processor and a commercial real-time multitasking kernel). Data collected using this platform shows that the intrusiveness can be quite significant.
A technique called “Routine Specialization” is proposed to lower that overhead. Results obtained from a “real-time-oriented” injector (RT-Xception) taken from the same system, show a very significant improvement. A comparison with data from other authors shows that with this change SWIFI becomes a viable technique for a wide range of real-time applications.
KeywordsFault Model Target System Fault Injection Transient Fault Permanent Fault
Unable to display preview. Download preview PDF.
- 1.Carreira, J., Madeira, H., Silva, J.G.: Xception: A Technique for the Experimental Evaluation of Dependability in Modern Computers. IEEE Trans. on Software Engineering, February (1998) 125–135Google Scholar
- 2.Chillarege, R., Bowen, N. S.: Understanding Large System Failures: A Fault Injection Experiment. FTCS-19, Chicago-IL (1989)Google Scholar
- 4.Czeck, E.: Estimates of the Abilities of Software-Implemented Fault-Injection to Represent Gate-Level Faults: IEEE Int’l Workshop on Fault and Error Injection for Dependability Validation of Computer Systems, Gothenburg-Sweden (1993)Google Scholar
- 5.Echtle, K., Leu, M.: The EFA Fault Injector for Fault-Tolerant Distributed System Testing. Workshop on Fault-Tolerant Parallel and Distributed Systems (1992)Google Scholar
- 6.Fuchs, E.: An Evaluation of the Error Detection Mechanisms in MARS using Software-Implemented Fault Injection. EDCC-2, Taormina-Italy (1996)Google Scholar
- 7.Gait, J.: Probe Effect. IEEE Trans. on Parallel and Distributed Systems (1992)Google Scholar
- 8.Han, S., Rosenberg, H. A., Shin, K. G.: DOCTOR: an Integrated Software Fault Injection Environment. IEEE Int’l Workshop on Integrating Error Models with Fault Injection, Annapolis-Maryland-USA (1994)Google Scholar
- 9.Hsueh, M.-C., Tsai, T. K., Iyer, R. K.: Fault Injection Techniques and Tools. IEEE Computer, April (1994) 75–82Google Scholar
- 10.Jenn, E., Arlat, J., Rimén, M., Ohlsson, J., Karlsson, J.: Fault Injection into VHDL Models: The MEFISTO Tool. FTCS-24, Austin-Texas-USA (1994)Google Scholar
- 13.Kao, W., Iyer, R. K., Tang, D.: DEFINE: A Distributed Fault Injection and Monitoring Environment. Workshop on Fault-Tolerant Parallel and Distributed Systems (1994)Google Scholar
- 14.Laprie, J.C.: Dependability: Basic Concepts and Terminology. Springer-Verlag (1991)Google Scholar
- 15.Lovric, T.: Processor Fault Simulation with ProFI. European Simulation Symposium (1995) 353–357Google Scholar
- 16.Muller, G., Marlet, R., Volanski, E. N., Consel, C., PU, C., Goel, A.: Fast, Optimized SUN RPC Using Automatic Program Specialization. 18th International Conference on Distributed Computing Systems, Amsterdam-The Netherlands, May (1998)Google Scholar
- 17.Powell, D, Veríssimo, P. Bonn, G., Waeselynck, F., Seaton., D.: The Delta-4 Approach to Dependability in Open Distributed Computing Systems. FTCS-18, Tokyo (1988)Google Scholar
- 18.Rimen, M., Ohlsson, J., Torin, J.: On Microprocessor Error Behaviour Modelling. FTCS-24, Austin-Texas (1994)Google Scholar
- 19.Rosenberg, H.A., Shin, K.G.: Software Fault Injection and its Application in Distributed Systems. FTCS-23, Toulouse (1993)Google Scholar
- 20.SMX® Simple Multitasking Executive, http://www.smxinfo.com
- 21.Segall, Z., Vrsalovic, D., Siewiorek, D., Yaskin, D., Kownacki, J., Barton, J., Dancey, R., Robinson, A., Lin, T: FIAT: Fault Injection Based Automated Testing Environment. FTCS-18, Tokyo (1988)Google Scholar
- 22.Silva, J.G., Prata, P., Rela, M., Madeira, H.: Practical Issues in the Use of ABFT and a New Failure Model. FTCS-28, Munich-Germany (1998)Google Scholar
- 23.Timmerman, M.: Is Windows CE 2.0 a real threat to the RTOS World?. Real-Time Magazine, vol. 98-3 (1998)Google Scholar
- 24.Tsai, T.K., Iyer, R.K., Jewitt, D.: An Approach towards Benchmarking of Fault-Tolerant Commercial Systems. FTCS-26, Sendai-Japan (1996)Google Scholar
- 25.Young, L.T.: A Hybrid Monitor Assisted Fault Injection Environment, DCCA-3, Sicily-Italy (1993)Google Scholar