Enable physical test of RFID physics in productive environment withpout affecting performance or throughput
Original Publication Date: 2005-Sep-26
Included in the Prior Art Database: 2005-Sep-26
Once a physical RFID Infrastructure is build and multiple components mounted together there is no way to figure out it works as designed. Works as designed means that an RFID read takes places, this read will be filtered and only for the business process necessary events will be uploaded. A decision will be done in the RFID Middleware or the legacy application and then a feedback will be generated and published close to the RFID reader. The idea is now to verify if all components are mounted together correctly without any upload into the legacy application.
Enable physical test of RFID physics in productive environment withpout affecting Enable physical test of RFID physics in productive environment withpout affectingEnable physical test of RFID physics in productive environment withpout affecting Enable physical test of RFID physics in productive environment withpout affecting
performance or throughput performance or throughputperformance or throughput performance or throughput
Any RFID implementation will be in general relative close to the IBM* Domain model . This is described and documented in the following pictures :
Overview: End-to-end component model of the RFID middleware; RFID Middleware is highlighted with blue box.
Systems Management Domain
Antenna & Reader Domain
Process Integration Domain
& Business Application
Object Directory Domain
Security & Data Privacy Management
Edge Domain Edge Domain
Motion Sensor Motion Sensor
I/O cable low voltage
Usually every component can be tested standalone (e.g. a light tree, a reader, the antenna). For complex installations there is no generic test software or
automated test system available since the number of possible installations is very huge. An integration test only is able to verify if all components work together as designed. This complex test cases are usually run as with a subset of the components as a black box test without having all components connected together . The problem after running this tests is, that there is no guarantee on a successful end-to-end scenario because physical enhancements are done after such a testcase has finished. Either a network cable will be moved from the test system to the production environment or a backend simulator will be replaced by the real backend or a reconfiguration of a test environment to become a production environment is done.
The idea is to have a test scenario within a fully plugged together and configured environment. Within the test there will be read a tag which is defined as "good". The antenna and the reader will detect the tag and upload its information into the next layer. All layers will be executed - without the Enterprise and Business Application Domain. The Business Process Integration Domain will decide that this tags is seen as a good tag and will return an appropriate answer. This answer then will lead to turn on a light, sound a beep or whatever was configured.
A similar testcase will be executed for a "bad" tag. Walkthrough of the proposed solution :
Standard Procedure (all messages will be send upwards)
(1) The µBroker engine will forward all messages into the standard workflow . No analysis on the tag format or tag data is done .
uplink µBroker to MQ
downlink MQ to µB...