Janez Golob (2009) Distributed control system: implementation options. EngD thesis.
Abstract
THE PURPOSE OF THIS BACHELOR WORK IS TO EXAMINE THE POSSIBILITY OF IMPLEMENTING A DISTRIBUTED CONTROL SYSTEM. IN THIS CASE IT WAS NECESSARY TO OBTAIN AND ANALYZE FUNCTIONAL AS WELL AS NON-FUNCTIONAL REQUIREMENTS THAT MUST BE SUPPORTED BY MIDDLEWARE. FIRSTLY I DETERMINED IF AND TO WHAT EXTENT WERE EACH OF MIDDLEWARE SOFTWARE NON-FUNCTIONAL REQUIREMENTS SUFFICIENT, BY SEARCHING THROUGH AVAILABLE LITERATURE, CONTACTS WITH MANUFACTURERS AND THE SOURCE CODE REVIEWS. THEN I DEVELOPED A SIMPLE CONTROL SYSTEM FOR EACH MIDDLEWARE SOFTWARE, TESTED IT ON A TEST ENVIRONMENT AND MEASURED PERFORMANCE. USING THE RESULTS OF MEASUREMENTS AND REQUIREMENTS STUDY, I COMPARED THE MIDDLEWARE SOFTWARES AND CATEGORIZED THEM FROM WORST TO BEST. IT TURNED OUT THAT NO OF THE INVESTIGATED TECHNOLOGIES IS A PERFECT fiT FOR THE REQUIREMENTS AND THAT EACH HAD ADVANTAGES AND DRAWBACKS. RTI DDS FOR EXAMPLE IS THE MOST POWERFUL MIDDLEWARE SOFTWARE, BUT DOES NOT OffER ANY CONTROL SYSTEM SPECIfiC TOOLS AND SERVICES. AS THE BEST COMPROMISE BETWEEN CAPACITY, SCALABILITY, THE NUMBER OF TOOLS AND SERVICES, ETC. HAS PROVEN EPICS.
Actions (login required)