Thoughts about OPC-UA

The major problem with OPC Classic is that it is based on COM and COM+ technology which requires the use of DCOM for remote connectivity. As anyone that has a PC and has to do security updates every other day knows there have allways been major security wholes in DCOM. It is also very hard to trust domains so that you can share data across them. Lastly OPC Classis has not built in encryption to protect the datt from sniffing.

With that said OPC did do what it was intended to do which was to get the industry standardized on a way to share data between applications thereby openging a multitude of offerongs for the end user. In short it was a very good business decission because it open every manufacturer of software's horizontal market offering with a minimum of effort. Minimum being a relative term here.

So what does OPC UA offer. first it encapsulates all the standars of OPC Classic. This means that proven technology can still be used. It allows that data be shared on using Web based services technology whch is proven to be reliable and secure. It allows for very high levels of encryption so that data is secure. Because of the new technoloby a companies HQ in the US can get live production data from a plant in South America with out woring about security.

So to summarize, yes I think it is a natural progression.

Fred Loveless

Applications Manager

Kepware Technologies

I think it is the natural progression. I think the foundation shot a little too high when trying to develop it- instead of simply creating a better version of classic OPC, they tried to create a much wider spec that would accommodate a larger range of scenarios. That's great in theory, but the reality is that it's really bogged down the development/initial adoption stage.

That said, it's really a great improvement over the current situation (first and foremost: platform independence), and once we start seeing more products roll out to market, I think it's really going to become the standard. The thing is, from a product point of view, there's a bit of a deadlock: server vendors don't feel pressure because there are no clients, client vendors don't feel enabled because there are no servers, and the foundation is kind of lagging behind on all fronts. This is starting to break up, though, with a few big UA products already released this year, and more to come I'm sure. By end of 2010, I suspect the momentum will have shifted in favor of UA.

I've been using OPC-UA for about 3-4 months now. Moved from Kepware to an OPC-UA server. I love the performance and the ease of setup, as well as the builtin security options. Currently subscribed to ~30,000 tags/second, and it has been solid. I could not get the performance I needed without clustering OPC server with OPC DA, OPC UA works well for my applications, but YMMV.