July 22nd, 2013 by
By Fabrice Boisset
General Manager, EMEA
Given the potential cost savings, performance boosts and scalability gains afforded by moving to an innovative RTX real-time platform, the inquiries about porting real-time applications to RTX are on the increase.
And it’s hardly surprising that the very first question we’re often asked when visiting prospects is about the time required to port an application to RTX.
“It depends,” isn’t a very satisfying response, but in fact there is no magic porting formula. It is impossible to provide an instant or precise answer because there are many parameters to consider in estimating the resources and time that are required.
The good news that we hear regularly from existing RTX customers is that porting a 32-bit RTX application to RTX64 is very straight forward. Having the same architecture, the same API and the same development environment (Microsoft Visual Studio) are contributing factors in this ease of transition.
It is when porting from a completely different architecture that many variables must be taken into consideration.
All real-time operating systems (RTOSes) do not have same architecture (multi-tasking), nor the same API (POSIX vs. Win32), nor the same mechanisms (available real time objects). As one example, let’s say you developed your existing real-time application using a message passing mechanism. In order to communicate between tasks you might need to adapt your code to use standard IPC mechanisms such as shared memory.
The size of the application is important to consider as well. The type of code and number of lines of code in your original application will also have a bearing on the port effort, as will the programming language.
There are models to estimate resources required to develop software. They take in consideration all the parameters mentioned here, and can be used for porting as well.
The economic and performance benefits of porting are well worth the time and effort.