@InProceedings{marsden02dependability,
  author       = {Marsden, Eric and Fabre, J-C and Arlat, Jean},
  booktitle    = {21st IEEE Symposium on Reliable Distributed Systems, 2002. Proceedings.},
  title        = {Dependability of CORBA systems: Service characterization by fault injection},
  organization = {IEEE},
  pages        = {276--285},
  comment      = {* analyze problem of characterizing middleware platforms

  * incl. injection at communication level
  * plus case study: CORBA naming service, CORBA event service

* dependability assessment techniques of middlewares

  * two approaches (xor):

    * model-based

      * stress design-time capabilities

    * measurement-based

      * stress higher accuracy
      * two main, complementary approaches

        * observe faults in large real-life systems

          * need "good-enough" monitoring
          * but failures are rare

        * inject faults into target system

          * provides results:

            * understanding of systems' failure modes
            * understanding of error detection and fault tolerance
              mechanisms

        * complementary:

          \#. derive fault model from field observations
          \#. do fault injection experiments according to fault model

* describe middleware architecture

  * \# CORBA-specific, not taking notes here

* fault model for CORBA-based systems

  * incl. physical, software, resource-management, communication faults

* failure modes of ORB

  * {crash, hang} of {object, process, thread}
  * corruption of (middleware's) IO
  * {omission, duplication of messages}
  * incorrect exception handling

* methods for fault injection

  * code mutation
  * memory corruption
  * robustness testing (stress through incorrect input)
  * syscall interposition
  * network: corruption, reordering, loss, duplication

* factors to consider for fault injection

  * fault model
  * workload
  * observations

* \# more elaborations on network fault injection
* plus case study: CORBA naming service, CORBA event service

  * crashes, hangs, and application exceptions for both experiments
  * suggests, that ORB is stressed instead of services themselves},
  file         = {:marsden02dependability - Dependability of CORBA systems_ Service characterization by fault injection.pdf:PDF},
  groups       = {fault injection},
  timestamp    = {2019-07-02},
  year         = {2002},
}

{"_id":"y46XS6vXG5Q96KZQN","bibbaseid":"marsden-fabre-arlat-dependabilityofcorbasystemsservicecharacterizationbyfaultinjection-2002","author_short":["Marsden, E.","Fabre, J.","Arlat, J."],"bibdata":{"bibtype":"inproceedings","type":"inproceedings","author":[{"propositions":[],"lastnames":["Marsden"],"firstnames":["Eric"],"suffixes":[]},{"propositions":[],"lastnames":["Fabre"],"firstnames":["J-C"],"suffixes":[]},{"propositions":[],"lastnames":["Arlat"],"firstnames":["Jean"],"suffixes":[]}],"booktitle":"21st IEEE Symposium on Reliable Distributed Systems, 2002. Proceedings.","title":"Dependability of CORBA systems: Service characterization by fault injection","organization":"IEEE","pages":"276–285","comment":"* analyze problem of characterizing middleware platforms * incl. injection at communication level * plus case study: CORBA naming service, CORBA event service * dependability assessment techniques of middlewares * two approaches (xor): * model-based * stress design-time capabilities * measurement-based * stress higher accuracy * two main, complementary approaches * observe faults in large real-life systems * need \"good-enough\" monitoring * but failures are rare * inject faults into target system * provides results: * understanding of systems' failure modes * understanding of error detection and fault tolerance mechanisms * complementary: #. derive fault model from field observations #. do fault injection experiments according to fault model * describe middleware architecture * # CORBA-specific, not taking notes here * fault model for CORBA-based systems * incl. physical, software, resource-management, communication faults * failure modes of ORB * crash, hang of object, process, thread * corruption of (middleware's) IO * omission, duplication of messages * incorrect exception handling * methods for fault injection * code mutation * memory corruption * robustness testing (stress through incorrect input) * syscall interposition * network: corruption, reordering, loss, duplication * factors to consider for fault injection * fault model * workload * observations * # more elaborations on network fault injection * plus case study: CORBA naming service, CORBA event service * crashes, hangs, and application exceptions for both experiments * suggests, that ORB is stressed instead of services themselves","file":":marsden02dependability - Dependability of CORBA systems_ Service characterization by fault injection.pdf:PDF","groups":"fault injection","timestamp":"2019-07-02","year":"2002","bibtex":"@InProceedings{marsden02dependability,\n author = {Marsden, Eric and Fabre, J-C and Arlat, Jean},\n booktitle = {21st IEEE Symposium on Reliable Distributed Systems, 2002. Proceedings.},\n title = {Dependability of CORBA systems: Service characterization by fault injection},\n organization = {IEEE},\n pages = {276--285},\n comment = {* analyze problem of characterizing middleware platforms\n\n * incl. injection at communication level\n * plus case study: CORBA naming service, CORBA event service\n\n* dependability assessment techniques of middlewares\n\n * two approaches (xor):\n\n * model-based\n\n * stress design-time capabilities\n\n * measurement-based\n\n * stress higher accuracy\n * two main, complementary approaches\n\n * observe faults in large real-life systems\n\n * need \"good-enough\" monitoring\n * but failures are rare\n\n * inject faults into target system\n\n * provides results:\n\n * understanding of systems' failure modes\n * understanding of error detection and fault tolerance\n mechanisms\n\n * complementary:\n\n \\#. derive fault model from field observations\n \\#. do fault injection experiments according to fault model\n\n* describe middleware architecture\n\n * \\# CORBA-specific, not taking notes here\n\n* fault model for CORBA-based systems\n\n * incl. physical, software, resource-management, communication faults\n\n* failure modes of ORB\n\n * {crash, hang} of {object, process, thread}\n * corruption of (middleware's) IO\n * {omission, duplication of messages}\n * incorrect exception handling\n\n* methods for fault injection\n\n * code mutation\n * memory corruption\n * robustness testing (stress through incorrect input)\n * syscall interposition\n * network: corruption, reordering, loss, duplication\n\n* factors to consider for fault injection\n\n * fault model\n * workload\n * observations\n\n* \\# more elaborations on network fault injection\n* plus case study: CORBA naming service, CORBA event service\n\n * crashes, hangs, and application exceptions for both experiments\n * suggests, that ORB is stressed instead of services themselves},\n file = {:marsden02dependability - Dependability of CORBA systems_ Service characterization by fault injection.pdf:PDF},\n groups = {fault injection},\n timestamp = {2019-07-02},\n year = {2002},\n}\n\n","author_short":["Marsden, E.","Fabre, J.","Arlat, J."],"key":"marsden02dependability","id":"marsden02dependability","bibbaseid":"marsden-fabre-arlat-dependabilityofcorbasystemsservicecharacterizationbyfaultinjection-2002","role":"author","urls":{},"metadata":{"authorlinks":{}},"downloads":0,"html":""},"bibtype":"inproceedings","biburl":"https://bibbase.org/network/files/AsPiHTmHHGjgy6xSQ","dataSources":["wjZw5s4JL49uLwn3p"],"keywords":[],"search_terms":["dependability","corba","systems","service","characterization","fault","injection","marsden","fabre","arlat"],"title":"Dependability of CORBA systems: Service characterization by fault injection","year":2002}