J
I just finished looking over the "draft" version of the ProfiNet specification from the PNO (Profibus Not-so Organization). I must say that it is remarkably unremarkable.
In order to avoid any conversational quagmires with respect to Industrial Ethernet as a whole, let us assume that the automation industry has a real need for Ethernet technologies visa-vi TCP/IP etc. What I see in ProfiNet is a rehashing of 5 year old technologies (at best) in a slightly different form.
Although I mostly agree with the overall vision of the integration of Engineering Applications (configuration tools, programming tools, runtime statistical analysis tools) into the Open Automation arena, the PNO clearly looses me from that point on. You really don't have to sweat the details to find some really glaring contradictions with the over all "Open Automation" objective. The most obvious one is the fact that ProfiNet engineering tools will use OLE/COM as the object model and they further state:
"The use of interface and communication standards that were developed in the Microsoft world does not necessarily mean that PROFInet is confined to Microsoft operating systems......"
then ONE paragraph later:
"....PROFInet Engineering Systems should therefore be conceived for the PC platforms with WINDOWS NT/2000 operating system."
Clearly an "Open Automation" system that is based on a closed proprietary operating system and that operating systems' technologies for distributed computing is less than ideal. Aren't we giving up a large amount of the benefits of "Open Automation". There are scores of purely Open, completely Committee based technologies for distributed computing based on component object principles. All of those technologies, besides being better object models by far, have implementations for Microsoft Windows platforms as well as dozens of other platforms. Sure, there are implementations of COM/DCOM that are not tied to Microsoft, but those only exist because the OPC foundation used COM/DCOM as their object model. At best, these are "catch up" copies of the original.
With all the choices for open interface, distributed computing technologies would COM/DCOM be chosen? Even Microsoft has abandon these technologies for enterprise business to business communications (because they really don't work in WAN implementations). Even SOAP would have been a better choice. But now that they have chosen COM/DCOM model, why create a new standard at all? Why not just use OPC?
Really what is the point here? Do they really want to make a truly modern, open and innovative automation platform or as quickly as possible pump out the same old technologies that have been kicking around for years thinly disguised as a NEW cutting edge technology?
In order to avoid any conversational quagmires with respect to Industrial Ethernet as a whole, let us assume that the automation industry has a real need for Ethernet technologies visa-vi TCP/IP etc. What I see in ProfiNet is a rehashing of 5 year old technologies (at best) in a slightly different form.
Although I mostly agree with the overall vision of the integration of Engineering Applications (configuration tools, programming tools, runtime statistical analysis tools) into the Open Automation arena, the PNO clearly looses me from that point on. You really don't have to sweat the details to find some really glaring contradictions with the over all "Open Automation" objective. The most obvious one is the fact that ProfiNet engineering tools will use OLE/COM as the object model and they further state:
"The use of interface and communication standards that were developed in the Microsoft world does not necessarily mean that PROFInet is confined to Microsoft operating systems......"
then ONE paragraph later:
"....PROFInet Engineering Systems should therefore be conceived for the PC platforms with WINDOWS NT/2000 operating system."
Clearly an "Open Automation" system that is based on a closed proprietary operating system and that operating systems' technologies for distributed computing is less than ideal. Aren't we giving up a large amount of the benefits of "Open Automation". There are scores of purely Open, completely Committee based technologies for distributed computing based on component object principles. All of those technologies, besides being better object models by far, have implementations for Microsoft Windows platforms as well as dozens of other platforms. Sure, there are implementations of COM/DCOM that are not tied to Microsoft, but those only exist because the OPC foundation used COM/DCOM as their object model. At best, these are "catch up" copies of the original.
With all the choices for open interface, distributed computing technologies would COM/DCOM be chosen? Even Microsoft has abandon these technologies for enterprise business to business communications (because they really don't work in WAN implementations). Even SOAP would have been a better choice. But now that they have chosen COM/DCOM model, why create a new standard at all? Why not just use OPC?
Really what is the point here? Do they really want to make a truly modern, open and innovative automation platform or as quickly as possible pump out the same old technologies that have been kicking around for years thinly disguised as a NEW cutting edge technology?