first of all, it is zz's. My view on the original text is that the specification of distributed system is not so easy to work out. Therefore, OPC-DA is irreplaceable for the time being. OPC-XMLDA is a new technology across firewalls, but its real-time performance is poor. The source of this article is a comment on another article criticizing opc. (The link of the original text is/sup _ zoux/9284/message.aspx # 3648, and the title of the article being commented is OPC-the morbid agreement between capital and worshipping foreign countries) zz article is as follows:

The landlord seems to be quite familiar with OPC, but he doesn't seem to know it very well.

First of all, OPC is a protocol, in which only the interface is defined. The disadvantage of OPC is that it is based on DCOM, and most of the criticisms come from DCOM itself rather than OPC protocol itself. At least, the dissatisfaction of this article with OPC almost comes from DCOM. Then the landlord should scold Microsoft, or the OPC foundation should compromise with Microsoft, not the OPC agreement itself. Does the landlord want to develop OPC protocol into a protocol like modbus, or does he propose to establish a distributed framework to replace DCOM? These two things are totally different from each other. If it is the former, I advise the landlord to stand on a higher level, instead of blindly preparing to do useless work. If it is the latter, I am quite interested, and I am only ignorant. The distributed framework only knows DCOM, Cobra, JMI,. Net Remoting and SOAP. Among them, JMI and. Net Remoting still depend on the platform at present, while Cobra looks beautiful, and SOAP seems to be the future development direction. At present, OPC 3. protocol is already based on SOAP, and its performance problems prevent it from being popularized on a large scale.

Secondly, OPC protocol has existed for many years, so it seems a bit excessive and petty to criticize an old thing from the present point of view. Just as we can't say what a crude IPC technology DDE is now, after all, it is the product of an era. In the era of OPC, isn't DCOM the mainstream technology on windows, unless it is separated from Microsoft from the beginning?

Third, the word "capital and worship of foreign countries" seems to be too cynical. If OPC is used as "capital and worship of foreign countries", then most of us who use computers are "capital and worship of foreign countries", because the windows, linux, unix and solaris we use are all given to us by foreigners. If one day it is true,

Finally, I gave the landlord a sentence, "Learn from foreigners and learn from them to govern foreigners". Before governing foreigners, I learned from foreigners rather than despised them.

The original text is as follows:

OPC-a morbid agreement between capital and worshipping foreign countries

Although most manufacturers support OPC protocol at present and regard it as an open standard. I have been engaged in real-time database research and development for a long time, and have in-depth research on OPC protocol. So far, apart from sadness, there is only one thing I have to say. Is OPC really advanced? For some colleagues who have been developing non-standard products by writing serial port protocols in the past, it seems that they have just felt its advantages, and writing some OPC interfaces for receiving projects, etc., may feel mysterious and unattainable. In fact, OPC is just a windows sockets definition based on Microsoft DCOM technology. In its design, it did not consider many hardware conditions necessary for industrial control, but just moved Microsoft DCOM technology to the industrial control field intact. In recent years, some colleagues have contacted SUPCON SOFT every year, hoping to solve the problem of OPC interface. As the first member of OPC Foundation and an advocate of domestic OPC Foundation, SUPCON knows OPC very well and has a large number of programmers who can develop OPC interfaces. But this does not mean that SUPCON will undertake the service of these interface problems. As an enterprise, its professionalism lies in providing its own professional products and services with core values, not others. But it also shows that Chinese people misunderstand and blindly follow OPC interface from another side. Is OPC really beautiful? Since its application, OPC has brought more pain than benefits. DCOM is a set of services that rely heavily on Microsoft technology. Only one OPC limits the diversity of operating systems in the industrial control field at present. This is nothing. If you are patriotic, China really doesn't have a remarkable operating system. But there are too many problems with OPC:

※ The security configuration is complicated: for industrial control personnel who are not professional in the operating system, the security configuration of OPC is too professional and complicated. This leads to many instances, OPC is not self-activated after the system starts, but requires interactive users to log in, which brings great insecurity to the system. That is, every system restart may require human intervention. Although it can be avoided through reasonable DCOM configuration, unfortunately, most industrial control practitioners do not grasp this;

※ Remote activation is difficult: If two computers are no longer in a "domain" with strong Microsoft technical characteristics, remote activation of OPC is a nightmare. In many projects, this configuration alone will make many engineers feel sad. Do you know how activation between different domains is done in most projects? Hehe, many colleagues chose two machines to log in with the same user name and password to undermine security; Other colleagues who master some programming skills save the user name and password of another computer in one computer; The reason why these security problems cannot be ruled out is the damn OPC protocol, which is a cancer adsorbed on Microsoft's DCOM technology;

※ Complex development: Although I have a good command of DCOM technology, I can still recall the dark days of learning DCOM programming when I was young. After a period of pain, DCOM is an epiphany, and it is found that all the people who wrote DCOM textbooks are mystifying and artificially increasing the complexity. At the same time, the memory management and calling technology of DCOM often requires more experience, which makes the communication development that was originally easy become anxious. That's why many colleagues in the industry entrust other companies to develop OPC interfaces.

※ Cross-platform difficulties: Even if you cross multiple operating systems of Microsoft, there will be some minor problems, and even fewer people can use OPC on Linux and UNIX; I'm only famous so far, and I've never seen such an expert in person;

So why did such a criticized protocol become the universally accepted standard today? There are two reasons, one is timing, and the other is capital. When the industrial Ethernet era was still in its infancy in HarmonyOS, and major automation manufacturers were still arguing endlessly about the future bus, Microsoft, the manufacturer of this operating system, used a distributed technology based on its own operating system to launch an unknown OLE for Process Control in an irrational era when DCOM seemed to be able to solve all distributed problems. It has not attracted enough attention from any automation manufacturer, and it is precisely because of this low-key admission, coupled with the usual conservatism of major automation manufacturers and short-sightedness about the development prospects of industrial Ethernet technology that OPC has grown up. Who will take an operating system manufacturer as a competitor? Therefore, the beginning of OPC is relatively smooth. Another powerful trumpeter is Microsoft. He didn't advocate OPC omnipotence, but it excessively advocated DCOM. Eventually, this capital operation brought impetuousness, and everyone simply stopped studying other open industrial Ethernet transmission protocols. OPC is a panacea. History repeats itself, and today we are blindfolded by the so-called SOA and Web2. technologies of IBM and other manufacturers.

Another reason is to worship foreign countries. Once upon a time, foreign things were extremely good. I still remember that an internal TCP-based data transmission protocol was defined at that time, and conservatives kept chattering in my ear: the protocol was made by large foreign companies, so how magical and professional it was. In short, China people didn't even have the ability to make a TCP transmission protocol for an enterprise. However, in the end, it is proved that not only can it be made, but China people can make an excellent open data transmission protocol as long as they have a good grasp of the demand for industrial data transmission. But the question always seems to arise: who supports you when you make it? You made, well, although it is an open agreement, why did your company A make it, not my company B? There are so many problems for Chinese people. At present, there are more than a dozen automation manufacturers with a certain scale in China. Have you set up a standards committee for many enterprises to discuss the state-owned open standards? No! That is the reality. Aren't we proud to be recognized by the American Instrument Association? Aren't we glad to be able to meet European standards? Therefore, in such soil, it is rare for native seeds to blossom and bear fruit.

actually, the most suitable Ethernet data exchange protocol for industrial use is definitely not OPC, but a TCP/IP-based, platform-independent transmission protocol. As long as the real-time value, historical value and active change notification are taken into account, batch data reading and writing and concurrent connection are considered, and the processing speed of different devices is considered, this protocol will become robust and practical. And our people are fully capable of making their own open agreements!

I deeply know that although the problem is obvious, tomorrow morning, I still have to accept this world full of foreign brands and standards. Although OPC is not good, I'm afraid it will be in great demand in the next five years. Although my strength is limited, I am fortunate to work in a national automation enterprise, and I can do my best bit by bit. I hope my colleagues in the domestic industry can reach an understanding, and one day, I can jointly promote the open industrial Ethernet real-time data transmission standard formulated by China people. At that time, this automated industry can become simple, efficient and convenient because of the open standards. Just leave Microsoft.