A device may identify a plurality of first values associated with network traffic of a label-switched path of a plurality of label-switched paths. The device may determine an adjustment policy based on the plurality of first values. The adjustment policy may include one or more factors associated with a plurality of second values. The plurality of second values may be determined based on the plurality of first values. The device may implement the adjustment policy in association with the label-switched path. A bandwidth reservation of the label-switched path may be adjusted based on the adjustment policy. The adjustment policy may be implemented for fewer than all of the plurality of label-switched paths.

A method, apparatus, and machine readable storage medium is disclosed for balancing loads among a plurality of virtual machines (VMs) from a central dispatcher, wherein the dispatcher receives data packets and maps the data packets to VMs selected from the plurality of VMs, using a weighted hash function, having an associated weighting for each VM and forwarding each packet to a VM accordingly, wherein a load balancer decrements a weighting for a VM, responsive to an indication of the load on the VM exceeding a first load threshold. Weightings can correspond to a number of bins associated with each VM. Weightings are adjusted in response to receiving invite and disinvite messages from the VMs, representing their respective loads.

Embodiments are directed to a distributed computing system comprising a plurality of compute nodes for providing resources to users and a hierarchy of two or more layers of controllers coupling the compute nodes to a user interface via a control plane, wherein at least one compute node receives a local application program interface (API) call from an application running on the at least compute node, the local API call causing the at least one compute node to configure a local resource without requiring commands from the control plane.

Aspects of the technology described herein allocate limited computing resources, such as available bandwidth and battery power, to transferring the most urgent and important data from a client device to an online service. Client devices have enormous amounts of information about the user's activities that could be communicated to the service at any given time. However, the wireless transfer of information uses available battery power and can consume a user's data plan. The technology described herein uses a model to determine how often information should be sent to a service. The model can also determine what information to send. Different models can be implemented in different scenarios. The different models can include different weighting that will produce different decisions given the same inputs.

A method and system for distributive flow control and bandwidth management in networks is disclosed. The method includes: providing multiple Internet Protocol (IP) Gateways (IPGWs) that each have a maximum send rate and one or more sessions with associated throughput criteria, wherein each IPGW performs flow control by limiting information flows by the respective maximum send rate and throughput criteria; providing multiple Code Rate Organizers (CROs) that each have a bandwidth capacity, wherein each CRO performs bandwidth allocation of its respective bandwidth capacity to one or more IPGWs of the multiple IPGWs; interconnecting the multiple IPGWs with the multiple CROs; and performing bandwidth management across the multiple CROs and IPGWs. In the method, an IPGW of the multiple IPGWs provides flow control across a plurality of the CROs of the multiple CROs, and a CRO of the multiple CROs allocates bandwidth to a plurality of the IPGWs of the multiple IPGWs.

A computer-implemented method routes service requests to services in a service framework provided by a plurality of hosts. The method comprises receiving a service request for a service in the service framework and discovering a plurality of candidate hosts that host the service. The plurality of candidate hosts are a subset of the plurality of hosts. The method further comprises selecting a candidate host from the plurality of candidate hosts based on measured latencies for the plurality of candidate hosts and routing the service request to the selected candidate host.

A method includes receiving, at a control node of a cloud computing network, a first enterprise policy specific to the first enterprise and a second enterprise policy specific to the second enterprise, and managing communications between at least one user device of the first enterprise and the at least one enterprise application hosted on behalf of the first enterprise based on the first enterprise policy. The method also includes managing communications between at least one user device of the second enterprise and the at least one enterprise application hosted on behalf of the second enterprise based on the second enterprise policy.

A method for determining a Quality of Service (QoS) policy can be based on requested bandwidth. The method may initially receive a connection request which includes a requested bandwidth that corresponds to an application. The method may then determine a policy for an application data flow associated with the application based on the connection request. A bandwidth designation, which is based on the requested bandwidth, may be assigned to the application data flow based on the determined policy. Finally, the policy and the bandwidth designation may be provided so that a bearer can be assigned.

A network management system can include multiple network interfaces. For example, the network management system can include a first network interface that can receive a stream of network packets associated with users. The network management system can include a second network interface for transmitting the received stream of network packets. The network management system can shape the stream of network packets before transmission. The network management system can assign the network packets into classes. The classes may have a configured hierarchical relationship. The classes may also have an operational hierarchy based on bandwidth usage during operation. The network management system can shape the stream of network packets based on operational hierarchy of classes and configured hierarchical relationship.

Resource provider specifications, characterizing computing resources of computing resource providers, are received. The reachability of each IP address included in the received specification is determined. An agent is deployed that is operable to determine the value of each of a set of metrics in the environment of the host at which the agent is deployed. The agent determines the value of each metric of the set of metrics in the environment of the relevant host, and communicates the determined values to one or more computing devices that validate whether the resources characterized by the communicated values are sufficient to provide the performance characterized by the received specification and that each ISP router complies with a predetermined policy. For each computing resource provider validated and determined to comprise an ISP router compliant with policy, the specified computing resources are added to a pool of resources for cloud computing.