Computer systems and methods for allocating bandwidth so that server computers can send data to a client computer without exceeding the available bandwidth between the server computers and the client computer, or the processing bandwidth or capacity of the client computer, are discussed herein. In an embodiment, a method comprises determining a first share of bandwidth that is to be available for a first computing device to send data to the client computer during a first future period of time; determining a first rate at which the first computing device is to send data to the client computer during the first future period of time; determining that the first rate is less than the first share of bandwidth that is to be available for the first computing device to send data to the client computer during the first future period of time by a first delta; receiving, from a second computing device among the plurality of network computing devices, a first request for additional rate allocation; sending, to the second computing device, a first rate allocation that is equal to or less than the first delta; sending data to the client computer during the first future period of time at a first actual rate that is less than or equal to the first rate minus the first rate allocation.