The use case is very simple, and unless fixed, will make using SIS very hard for us.
Say you have 1M meters connected to the system, and a snow/ice storm hits (yes, I know you are in CA and never heard of such a thing :)). 100K meters all report a power outage. Due to the large number of meters reporting the problem, it may take the communication infrastructure 30 minutes to send this data in. But I would like to process the data as it flows in.
So I need to "pump" the system with CTIs.
But the timestamps on all the events is the same (they all suffered the outage at the exact same second). So I cannot send in any CTIs until I am sure all events got in, or some will get rejected. So maybe I wait an hour before I send in my CTI, to pump all the events out of SIS.
But wait, grandma is sitting in the dark, waiting for the utility to fix the power, but we are still waiting for a CTI.
This is all solved if we allow each event to be a point event, with its time set to DateTime.Now(), and having the original timestamp in the payload. We can set the CTIs to fire on every event, providing with a very responsive systems. We can then send these events to a appropriate systems to process.