Method for call admission control in MPLS networks

A network device, comprising:

• a memory configured to store instructions;
• and
• a processor configured to execute instructions in the memory to:
• obtain network topology information for a Multiprotocol Label Switching (MPLS) network;
• receive, from a services control device, a request with a flow demand for migrating a set of flows from a current data center to a selected data center in the MPLS network, wherein the request includes a flow size, an average packet size, an expected duration, and a list of endpoint locations;
• evaluate, based on the network topology information, the flow demand for the selected data center to determine if the selected data center can support the flow demand,
• wherein the evaluating includes examining local area network resources at either end of a link between the selected data center and a link endpoint for capacity to support specifications in the flow demand,
• wherein the local area network resources include remaining free subnet address space and Data Center Bridging (DCB)/Priority Flow Control (PFC) capabilities, and
• wherein the evaluating further includes modeling flows, based on the flow demand, for link aggregation groups (LAGs) between the selected data center and multiple endpoints identified in the list of endpoint locations;
• and
• send, to a network layer control device, a network soft-state to support the flow demand at the selected data center when the selected data center can support the flow demand, wherein the network soft-state represents a set of available network resources reserved to support the flow demand prior to admitting a flow corresponding to the flow demand.

A non-transitory computer-readable medium, including instructions executable by at least one processor, the computer-readable medium comprising one or more instructions for:

• receiving, from a service layer control device, a request including a flow demand for services at a selected data center of a Multiprotocol Label Switching (MPLS) network,
• wherein the request includes a request to mi rate a flow from a current data center of multiple data centers to the selected data center, and
• wherein the request includes a flow size, an average packet size, an expected duration, and a list of endpoint locations;
• first evaluating the flow demand for the selected data center to determine if the selected data center can support the flow demand,
• wherein the first evaluating includes examining first local area network resources at either end of a link between the selected data center and a link endpoint for capacity to support specifications in the flow demand, wherein the first local area network resources include remaining free subnet address space and Data Center Bridging (DCB)/Priority Flow Control (PFC) capabilities;
• determining, based on the evaluating of the selected data center, that the selected data center cannot support the flow demand;
• identifying an alternate data center of the MPLS network;
• second evaluating the flow demand against the alternate data center to determine if the alternate data center can support the flow demand, wherein the second evaluating includes examining second local area network resources at either end of a link between the alternate data center and a link endpoint for capacity to support the specifications in the flow demand, and wherein the second local area network resources include remaining free subnet address space and DCB/PFC capabilities;
• and
• sending, to a network layer control device, a network soft-state to support the flow demand at the alternate data center when the alternate data center can support the flow demand, wherein the network soft-state represents a set of available network resources reserved to support the flow demand prior to admitting a flow corresponding to the flow demand.

A method, comprising:

• collecting, by one or more computing devices, real-time network topology information for a Multiprotocol Label Switching (MPLS) network that includes multiple data centers for a content delivery system or a cloud-based computing service;
• receiving, by the one or more computing devices and from a service layer control device, a request including a flow demand for migrating a set of flows from a current data center to a selected data center of the multiple data centers,
• wherein the request including a flow demand is a request to migrate the flow from the current data center of the multiple data centers to the selected data center, and wherein the request includes a flow size, an average packet size, quality of service (QoS) handling requirements, an expected duration, and a list of endpoint locations serviced by the current data center;
• evaluating, by the one or more computing devices and based on the collected network topology information, the flow demand for the selected data center to determine if the selected data center can support the flow demand,
• wherein the evaluating includes examining local area network resources at either end of a link between the selected data center and a link endpoint for capacity to support specifications in the flow demand, and
• wherein the local area network resources include remaining free subnet address space and Data Center Bridging (DCB)/Priority Flow Control (PFC) capabilities;
• and
• sending, by the one or more computing devices and to a network layer control device, a network soft-state to support the flow demand at the selected data center when the selected data center can support the flow demand, wherein the network soft-state represents a set of available network resources reserved to support the flow demand prior to admitting a flow corresponding to the flow demand.