Should a DCS system server be used or not?
Thread Content
This post was last edited by zlky2005 on 2012-6-5 09:49. Currently, Shanghai Xinhua, HoliTech, and Xinhangzhi are approaching me to promote their DCS systems; I am working on the technical specifications. I would like to know whether a DCS system requires a server or not. For my DCS system, I have installed a 6KVA UPS power supply – is this sufficient? So, what is the network architecture for systems like Rohde & Schwarz’s MacSV, Xinhua’s XDP3000 system, and Guodian Zhishen’s systems?
Whether a DCS requires a server is related to its software architecture; in the case of a C/S architecture, an independent server must be configured. Generally, the engineer station can also serve as a configuration server, after all, configurations are not modified all the time. The C/S architecture data servers should be configured separately and redundantly; engineer stations or operator stations should not be used as servers. The capacity of the UPS is determined based on the system scale and power load, to ensure that a power outage does not disrupt the operation of the DCS; generally, the manufacturers of complete systems will carry out the calculations.
Whether a server is required for DCS depends on the network architecture of the DCS network. In my opinion, the server, acting as a communication buffer, can easily become a bottleneck in situations involving large amounts of data exchange. Online modification is not possible; (saying it is possible is a lie)
It is not recommended to use small systems. Large systems should have dedicated servers; they should not be shared with engineer workstations or similar
If it is the ePKS from American HONEYWELL, then a server is absolutely necessary. As for domestic DCS systems, as well as DCS systems from Japanese Yokogawa, American ABB, Siemens, and even Foxboro in the United States, I believe there is no need to install a server in such cases
A peer-to-peer network is not needed; a C/S network structure is essential.
So, what is the network architecture for systems like Rohde & Schwarz’s MacSV, Xinhua’s XDP3000 system, and Guodian Zhishen’s systems?
Still don’t understand? What are their structures respectively!
As users and maintainers, it would be best to avoid using servers, but many systems require servers to be set up – it depends on your choice!
Hilscher’s MACSV adopts a C/S network architecture, and it requires servers with 1:1 redundancy. It is advisable not to use engineer stations as servers, in order to avoid network transmission errors when modifying configurations; however, servers can certainly be used as operation stations.
Do the Advantech MacSV system, the Xinhua XDP3000 system, Guodian ZhiShen’s systems, and so on, all have the same network architecture? I’d like to know too; let’s discuss what the structure of each of our companies’ systems is.