160 likes | 258 Views
This tool injects faults in service data and messages, simulating mismatches, typos, format changes, conflicts, delays, and missing data to test system responses. It offers functionalities like static and dynamic changes to databases, perturbation of data items, and perturbation of SOAP messages. Developed at Politecnico di Milano, Italy, it generates detailed logs of applied data changes. Future work includes integrating QoS simulation, SH-BPEL, fault detection, and web-based interaction.
E N D
Torino, January 25th, 2007 WS-Diamond Meeting
A Data Fault Generator WP3 Team Politecnico di Milano, Italy
Data faults Generator • Data Disturber: Tool to inject faults on service data and messages • Treated failures • value mismatch • Typos (Jhon instead of John) • Different format (date) • Conflict in data values (London / Italy) • Delay in update operations • missing data
Data faults Generator • Includes 4 tools corresponding to 4 functionalities • Static changes to DB • Dynamic changes to DB • Dynamic data changes • Dynamic SOAP messages changes • Generates a detailed log about the applied data changes Milan, November 2006
…from WP3 Meeting in Milan… • Static (a priori) changes to DB • Dynamic changes to DB • Dynamic data changes • A set of methods (e.g., Set_To_Zero) to perturb single data items or databases • Methods can be collected in configurable operations (script) dynamically invoked in the service code to perform data changes or DB changes
Changes to DB Changes to local DB WS-Diamond node Non WS-Diamond node WS WS Local DB Local DB Network WS-Diamond node WS Simulator Local DB Global DB
Changes to Data WS-Diamond node Non WS-Diamond node WS WS Local DB Local DB Changes to internal service data Network WS-Diamond node WS Simulator Local DB Global DB
Dynamic SOAP messages changes • 2 Steps • Select which message and data should be perturbed • An handler intercepts each message and modifies the transported data
Step 1: Fault Configuration • Choose the service, operation and data to be perturbed • Choose the fault type and its statistical parameters Configuration record
Step 2: Handler • An handler captures each SOAP message • It checks if a perturb-operation related to the message exists • If exists the involved data are perturbed using our algorithms • A log stores each data change
Handler Service 1 Service 2 SOAP Message Perturb Handler Operations on Messages Perturb 1 ... Perturb N Log Management interface
SOAP messages Log White meat ItemID = O03 Seafood salad ItemID = O30
Future work • Expand the proposed architecture • Not only data fault simulator but also QoS simulator (data quality, performance, security…) • Integration with SH-BPEL • Inject faults in messages • Detect the faults presence (with logging tool) • Use SH-BPEL features to repear faulted services • Develop a web-based interface to interact with the simulator