AnsweredAssumed Answered

Question about how Cloud Connect works and about where the "Node" should sit?

Question asked by KPSyama on Aug 26, 2016
Latest reply on Aug 27, 2016 by Bryan Owen


Im a long time basic PI processbook and datalink user. I have recently been tasked with a project to share process data within our company and with our equipment suppliers. A bit about us. We are a mining company, with a head office in Perth, Australia. The mine site i work on is in Mali, Africa. Our internet connection out here is very limited, at best we have a 7MB connection for the whole site, and its very glitchy, drops out at least once a week or is restricted quite often. So onto my questions.


1) Can someone please help me understand how the cloud connect data flow works? I was led to believe that transferring data to "the cloud" was a one way process and then that data could be sent to multiple end users. A reason why i think this is good is the mine site is only uploading one set of information, then that information is being downloaded from "the cloud" by multiple end users. This is good for us in that the mine site is not sending data to one person, then when a second person is requesting information (potentially the same information) we are not using our bandwidth again to send same information to someone else. By sending data once to "the cloud" multiple peopel can then download directly from the cloud. This was my (very non-IT background) understanding of the process. Recent information i have read however paints a different picture. Apparently "the cloud" link is actually between the two nodes. So my node that i install on site will connect to the node the subscriber installs on their machine. Nos when i have multpile subscribers i'm still uploading information twice and hence not really saving any bandwidth. Can someoen please explain how the data flow works exactly?


2) I am still waiting (more than a week) for a cloud services account. That's a different story again. However, where would one suggest to install the node? Can the node be installed on the PI server itself? Or is it best to install the node on a virtual server separate to the PI Server? Also, there is a network link between the mine site and the head office. Would it be better to isntall the node on the Head office side of the network (which can still access the PI server) and then the node can use the internet connection from the Perth office to publish the data to all the end users?


Any feedback on helping me understand how this system works is greatly appreciated.