I would probably favor some sort of tunneling myself, either MAtrikon-style, or using locating the AIM OPC server on the machine with the client and using AIM*API to traverse the network between that box and the I/A as Alex suggested. I'll have to read that document you mention. Since the UA flavor uses SOAP/XML, there's probably more protocol overhead than even with DCOM, but the counter-argument is that machines and networks have gotten faster and can handle it (sounds like MS and Vista...). But I really know little about OPC UA. I still can't get past the feeling, though, that all this is overkill for most of the live data-transfer applications I use (historical data/A&E is another topic, but then "real-time" performance is not such an issue there). My gut tells me that something like Modbus/TCP or Ethernet/IP would probably do the job in 99% of cases. The extra features of the OPC flavors might be more beneficial in a MES/enterprise data integration environment, but for an APC, or foreign device integration, I'll take a simple protocol like a Modbus or CIP any day. A side note, and something I think about but don't dwell on because I don't think there is much I can do about it, is application-layer security. I doubt any vendor does a whole lot of security auditing or testing on the software that is used to implement clients and servers for these protocols, and those programs often run with root/admin privileges, making the host machine (i.e., your DCS) potentially vulnerable to remote exploits. Since simple protocols are easier to implement properly, and open protocols make it at least posssible for someone to implement an application-layer firewall, I would think simple, open protocols (like Modbus/TCP and Ethernet/IP) would eventually help lead to better security. Corey Clingo BASF Corporation <tom.vandewater@xxxxxxxxxxxxxx> Sent by: foxboro-bounce@xxxxxxxxxxxxx 05/02/2007 01:54 PM Please respond to foxboro@xxxxxxxxxxxxx To <foxboro@xxxxxxxxxxxxx> cc Subject Re: [foxboro] OPC again... I've heard people talk about the tunneler. To my understanding it uses TCP/IP to pipe from one machine to another instead of DCOM in a similar manner as Modbus TCP. I would feel a lot better about that than using DCOM. The old OPC DA,(Data Access),OPC HDA,(Historical Data Access),and OPC A&E (Alarms and Events), used COM/DCOM for its transport mechanism and it has been a good idea gone bad because of DCOM. There were several reasons that motivated the OPC Foundation to develop the new OPC UA, (Unified Architecture) and the limitations and problems associated with COM/DCOM were significant drivers for change. OPC UA based on Web Services/Service Oriented Architecture may alleviate a lot of these issues. The best article I've read to gain a better understanding of UA and its features and intent was written by Stefan-Helmut Leitner and Wolfgang Mahnke of the ABB Corporate Research Center in Germany. The link to the document, listed below will be broken and you will have to copy and paste it into your browser to make it work: http://pi.informatik.uni-siegen.de/stt/26_4/01_Fachgruppenberichte/ORA20 06/07_leitner-final.pdf Abstract: OPC Unified Architecture (OPC UA) is the new standard of the OPC Foundation providing interoperability in process automation and beyond. By defining abstract services, OPC UA provides a service-oriented architecture (SOA) for industrial applications - from factory floor devices to enterprise applications. OPC UA integrates the different flavors of the former OPC specifications into a unified address space accessible with a single set of services. This paper gives an overview of the architecture of OPC UA, its address space model and its services. It discusses the necessary security mechanisms needed to allow secure access over the internet. Finally, migration strategies to OPC UA applications are introduced.=20 Cheers, Tom VandeWater Control Systems Developer/Analyst Dow Corning Corporation Carrollton, KY USA _______________________________________________________________________ This mailing list is neither sponsored nor endorsed by Invensys Process Systems (formerly The Foxboro Company). Use the info you obtain here at your own risks. Read http://www.thecassandraproject.org/disclaimer.html foxboro mailing list: //www.freelists.org/list/foxboro to subscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=join to unsubscribe: mailto:foxboro-request@xxxxxxxxxxxxx?subject=leave