A: HTTP, TCP, and DNS requests are supported for Rate Limiting.
A: Yes. All limits are enforced through a designated action such as DROP, RESET, REDIRECT, or CACHE, when the limit or threshold is reached. Therefore, rate limiting feature takes an action to enforce the limit.
For example, if you apply rate limiting on an HTTP virtual server and limit it to 200K per client connection with a policy or limit identifier using "client.ip.src" with the "maxbandwidth" option, then the policy invokes the designated action on all subsequent requests when a connection reaches 200K.
In other words, if DROP is configured as the action, then all subsequent requests are dropped. If REDIRECT is configured as the action, then the subsequent requests are redirected to an alternate location as long as the limit is exceeded.
A: NetScaler Rate Limiting does not currently have the ability to control bandwidth without taking a designated action. Therefore an external network device such as a Packet Shaper is required to govern bandwidth utilization in this manner.
A: Rate Limiting can be used to limit a subset of clients using subnet or IP address to restrict them to the entity based on bandwidth as described. The rate limiting feature will not be applied to the clients outside that range.
Another option would be to utilize Content Switching feature to route high bandwidth consumers to a separate Load Balancing virtual server, such that it does not impact other client connections.