Weighted least connections divides active connections by server weight. The server with the lowest ratio gets the next request.
Server A: connections, weight , ratio = Server B: connections, weight , ratio =
Server A gets the next request because its ratio () is lower than B's ().
This algorithm combines capacity awareness with real-time load awareness. Time complexity is per request. Space is for connection counts and weights.
Use this when server capacities differ and request durations vary.