Method and system for allocating computation resources for latency sensitive services over a communication network

Systems and methods are provided for allocating computation capability. A network compute orchestrator (NCO) receives a request for service from a requester device. The NCO determines a required compute grade requirement for the service, and a maximum acceptable latency for the service. The NCO accesses a data structure that stores for each of a network arrangement of network devices: a number and type of available compute units, and latency to the requester device. The NCO selects a network device of the network arrangement of network devices that has an available compute unit meeting the required compute grade requirement, and has latency to the requester device that is lower than the maximum acceptable latency for the service. The NCO causes the selected network device of the network arrangement of network devices to perform the requested service for the requester device.