I could have sworn I mentioned this before, but I couldn’t find the thread here, so I think maybe I didn’t ;-)
For the past months have I been embedding IoTDB into applications. These applications allow configuring IoTDB at runtime and handle all sorts of special cases. All of these are “standalone” instances with one config node and one data node … all running inside the same VM. The one thing I was always quite unhappy about, is that even though everything is running inside one VM, I still needed to use the Session API, which makes absolutely no sense to me. Here I have strongly typed data, that I need to serialize send via socket to the same VM and on the other side parse it again. In most cases this doesn’t matter, but as some of you might have read on LinkedIn am I currently working on making IoTDB work as a module in Ignition (MES System). Right now, I’m still using an externally run IoTDB instance, so using the session API is the way to go, but I am planning on also providing an embedded version, where we can run IoTDB inside Ignition. Given sometimes very extreme volume of requests sent to IoTDB, I would really love to reduce the amount of wasted resources. So, what would be needed to have an Inner-VM API, that allows me to directly insert data into a data-node? Chris
