To better explain, we first need to know your application. That makes a huge difference. Here is an example on something I did a while ago. A sorting application, which meant the data was very important. It wasn't that high speed, and we weren't talking tons of data. Each sort was pushed into a FIFO with time stamp. PLC time synced to Ignition server. In a perfect world, the data would enter the FIFO and then immediately get pushed to SQL via transaction group. If something was broken, the data would sit there until able to send again. The sorting system would stop if the buffer got close to full, that ensured no data loss. But, that's just one example. As stated earlier, a ton of data and this would not be your best avenue.