Hello Aimin. Sorry to read of your problems. I'm trying to understand the scenario. Not sure if you really meant to write 3-400 customers. That would not constitute a large outage. Did you mean to write 4000? For various utilities we have tested much larger outages - even up to 100,000 in one hour - without issue. Can you duplicate the problem in your test environment (if you have one)? Or is it only occurring in production? Have you actually detected that the network becomes saturated in this scenario?
Thank you Dave for taking my question.
It is the production not dev, I don’t have dev RXIF set up yet.
And 3-400 Calls sometimes does trigger the issue, but the more customer/Call the longer interruption will presents.
I can recreate by sending a few hundreds of outages by jmeter.
I am not sure if the network become saturated or not , but our entire network get interrupted, include our phone system.
We are trying to get some help from Network gear vendor but I notice once responder server/services shut down, the interruption stops.
It might be a good idea to have the networking hardware monitored and traffic monitored from the switches to determine what type of traffic it's struggling with; and see if there is a configuration change the vendor can make to handle it better. Responder does generate a lot of traffic with it's design, but it's never been a problem for any type hardware vendor already in place with clients for past implementations. Some clients have created VLAN's or other methods of isolation to limit the impact of other users and maybe it's something that can be implemented here as well. Enabling of IGMP snooping can also limit the amount of multicast traffic being forwarded and should be enabled by the vendor too.