TPS stands for Transaction per Second. I am not sure if there is a best practice between the two I am in the process of implementing TPS based protection on my critical applications. The parameters are really going to be subject to your application and infrastructure, you need to look at what the TPS is currently and historically for the applications you are planning to protect, so you can determine what is normal and what is peak traffic for said application. One other thing you could look at would be at what point does the application or infrastructure tip over and set your threasholds somewhere between that and your historical high water mark for peak traffic.