- Newest
- Most votes
- Most comments
A single /64 has more than enough capacity for 'millions' of IPs; 18,446,744,073,709,551,616 to be specific. Can you expand on what kind of test you are running that needs to use more then that? What aspect of the stack are you testing that requires that many distinct client ips?
Hi, yes, however, the requirement that we have is for millions of /64 prefixes to be modelled and tested.
We want to set up and test a representative system; the devices in question control themselves the interface-id component, and in this case, those devices (not under our control) use a single interface-id.
SO we are in the situation of having millions of /64 prefixes allocated to those millions of devices, with each device using precisely one address - profligate.
IDEALLY we will not need to adjust the platform in order to be able to test; if we have to, then we CAN adjust the platform such that if we are in test mode, then we use interface-id to simulate load.
What aspect needs the millions of distinct client IPs - well, all aspects - we need to verify that accessing the data store performs under load for instance, as well as all of the network routing.
Relevant content
- asked 9 months ago
- asked 6 months ago
- asked 4 months ago
- asked 10 months ago
- AWS OFFICIALUpdated 7 months ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago
- AWS OFFICIALUpdated 2 years ago