Print

[leftwing47]

Hi,

I'm a JPPF newbie, but have had some grid experience in the past. Do you have any tips or suggestions on rate-limiting requests to nodes? I've a use case that requires quite strict control over how many requests are sent from a particular node IP address over a period of time. Basically, let's say I wanted to send 1 http request through a particular node IP address with a minimum 5 second gap?

I understand the load balancing angle, so assume this would be managed somehow through that, but wondered if anyone else has come across this kind of requirement?

Congratulations though: JPPF looks like a great project and I'm looking forward to investigating its potential!

[lolo]

Hello,

The load-balancing merely determines how many tasks from a job can be sent at once to a node, thus I believe it is way too coarse for your requirement. It seems to me that the rate-limiting would best be done in your own code and invoked explicitely from your tasks. For example, you could implement a service based on an executor that will queue the HTTP requests and execute them at the desired rate, even when multiple tasks execute in parallel in the same node. This doesn't require any specific configuration of the grid and provides the most flexibility.

Sincerely,
-Laurent

[leftwing47]

Thanks for the reply Laurent - very helpful!