Simplified frontend for data generation and testing purposes in WINCC OA
| Main Authors: | Puri, Urishita, Guzik, Jakub, Kulaga, Rafal |
|---|---|
| Format: | Report Journal |
| Bahasa: | eng |
| Terbitan: |
, 2017
|
| Subjects: | |
| Online Access: |
https://zenodo.org/record/1035173 |
Daftar Isi:
- Future NextGen Archiver, a research and development project of WinCC OA, is a joint collaboration of ETM Siemens and CERN. It supports multiple backends including a BigData backend using Apache Kudu as storage. For the completion of this project, a simplified frontend was designed which communicates via an external interface using ZeroMQ and Protocol Buffers (ProtoBuff) with the BigData backend. The data generator and tester are designed as plugins to independently use this interface to connect to the backend. The main purpose for data generator is to generate data that can be used for performance tests for different retrieval scenarios in this BigData backend. These scenarios include getting value changes over a period for a few signals specified by name or alias. It can be used to get value changes over a specified duration for many signals (possibly hundreds of thousands) specified by complex filters on metadata fields (like device name, location, type). Further, it can also be used to get values of all signals at start time and then replay value changes till the end time. The data generator was efficiently implemented using sorting and accumulation algorithms. As a result new values are produced by it at a rate such that data generation is not a bottleneck when compared with later stages of processing. The backend functional tester uses google tests in C++ 11 to perform functional testing of Kudu backend. It tests for response of backend under varied conditions of creation, modification and deletion of metadata and creation/duplication of data. As a result, functional testing for the backend and environment setup (data generation) for non-functional/ performance testing of the backend were favourably completed.