As i'm more and more drawn into the topic, as Cloud is maybe the biggest game changer in IT this decade, i am questioning how to cater with the demands we get. To make life easier, let's define 'Cloud' by MS Azure, be it IaaS (VM's) or PaaS (Azure framework). I feel using Azure and PI together is too much of a struggle. Customers complain that IT says everything is possible, but when it comes to deliver on that promise, there is a lot of technology that is just not mature enough.
Just some random topics i am looking at:
- Connections: I'm running some development servers on Azure VM's. Very handy, but one of the issues i run into is that clients don't connect properly to my Azure AF (no issues so far with Data Archive) due to KB01418 - PI AF Client Connection Issues to PI AF Server Hosted in Microsoft Azure. So no running DataLink, ProcessBook (not a big issue) or PSE or PI Builder (more of an issue) locally without issues. So i need to use remote desktop and spin up another VM to just host my clients. Not an expert on WCF, but can't we have an option to switch the connection to non-binary to workaround the issue? Anyhow, either you are in the cloud or out, nothing in between.
- AF SQL: When deploying my Development and Test environments on Azure i wanted to use the Azure SQL to run the PIFD DataBase. That does not work (not supported, i know) as the replication options in AF are not supported on Azure SQL (AF on Azure SQL ). If we would have a version of AF that does not support these requirements but does support Azure SQL we would have this option. Now we need licenses for SQL in our MS VM which is not what you like to do when you're using IaaS. For Development and Test replication and failover not really needed, and also not for most small-scale AF deployments.
- General: To get data in and out of my PI server in the cloud, we can of course use VPN tunnels and the like but that's taking steps back. We can use UFL, Web API, PI2PI, etc. But looking forward i would like to see some convergence on the patters to bridge the gap between on-premise and cloud. Standard ways to connect Azure IoT to On-Premise or cloud-based PI, etc.
- Pi Connect: Pi Connect works great, but it's been silent since then. I hoped for more capabilities to not only use the cloud to sync between on-prem PI systems, but also the hybrid variants. A mashup between PI Connector for BI and PI Connect?
- OMF (The OSIsoft Message Format v0.9 — OMF 0.9.0.0 documentation ): Looks like one way to have an integration between cloud and on-prem, but where does this tie into the roadmap? If i need to design something that is cloud-ready or cloud-compatible for the next 5 years what technology should i use to integrate with PI? Pushing / pulling data to / from PI in cloud or on-prem?
- Azure SQL for BI: Really great the Integrator for BI. But not having support for Azure SQL as a destination stumps me. That would be my #1 for anything on Azure.
Any thoughts and ideas welcome! Better still, new releases or plans...
PS: not saying i don't like PI, it is a turbulent era for all of us in IT. I'm just trying to see a way to cut through the pattern of finding solutions that work only 90%.