Data network controller

An apparatus for controlling a data flow on a network, wherein the network includes a plurality of network service providers (NSPs) and the data flow is currently routed through a current NSP, the apparatus comprising at least one computing device configured to:

• evaluate a candidate latency for a candidate NSP, wherein the candidate NSP is selected from remaining ones of the plurality of NSPs;
• evaluate a candidate loss for the candidate NSP;
• evaluate a candidate bandwidth for the candidate NSP to a bandwidth allocation for the candidate NSP;
• determine a loss value for the remaining ones of the plurality of NSPs;
• determine a cost for the remaining ones of the plurality of NSPs;
• evaluate an identifier for the remaining ones of the plurality of NSPs;
• and
• determine a change value.

An apparatus for controlling a data flow on a network, wherein the network includes a plurality of network service providers (NSPs) and the data flow is currently routed through a current NSP, the apparatus comprising at least one computing device configured to:

• compare a candidate loss for a candidate NSP to a current loss for the current NSP, wherein the candidate NSP is selected from the remaining ones of the plurality of NSPs;
• compare a candidate latency for the candidate NSP to a current latency for the current NSP;
• determine a best latency of the remaining ones of the plurality of NSPs;
• determine a best loss of the remaining ones of the plurality of NSPs;
• evaluate a candidate bandwidth for the candidate NSP;
• determine a maximum candidate bandwidth value for the candidate NSP;
• determine a cost tier for the remaining ones of the plurality of NSPs;
• determine a lowest identifier of the remaining ones of the plurality of NSPs;
• and
• assign a change value to control routing of the data flow.

A computer-readable medium storing instructions for controlling data flow on a network, wherein the network includes a plurality of network service providers (NSPs) and the data flow is currently routed through a current NSP, by performing the steps of:

• evaluating a candidate latency for a candidate NSP, wherein the candidate NSP is selected from remaining ones of the plurality of NSPs;
• evaluating a candidate loss for a candidate NSP;
• comparing a candidate bandwidth for a candidate NSP to a bandwidth allocation for a candidate NSP;
• determining a loss value for at least some of the remaining ones of the plurality of NSPs;
• determining a cost for at least some of the remaining ones of the plurality of NSPs;
• evaluating an identifier for at least some of the remaining ones of the plurality of NSPs;
• and
• determining a change value.

A computer-readable medium storing instructions for controlling a data flow on a network, wherein the network includes a plurality of network service providers (NSPs) and the data flow is currently routed through a current NSP, by performing the steps of:

• comparing a candidate loss for a candidate NSP to a current loss for the current NSP, wherein the candidate NSP is selected from remaining ones of the plurality of NSPs;
• comparing a candidate latency for the candidate NSP to a current latency for the current NSP;
• determining a best latency for the remaining ones of the plurality of NSPs;
• determining a best loss for the remaining ones of the plurality of NSPs;
• evaluating a candidate bandwidth values for the candidate NSP;
• determining a maximum candidate bandwidth value for the candidate NSP;
• determining a cost tier for the remaining ones of the plurality of NSPs;
• determining a lowest identifier for the remaining ones of the plurality of NSPs;
• and
• assigning a change value to control the data flow.

A method for distributing data flows between a plurality of network system providers (NSPs), wherein each NSP has a multi-tier cost structure and a minimum commitment level, the method comprising:

• providing a data flow on a current NSP;
• for a candidate NSP, wherein the candidate NSP is one of the remaining NSPs:
• comparing performance data for the current NSP with performance data for the candidate NSP;
• if the performance data for the candidate NSP is better than the performance data for the current NSP, then determining whether current bandwidth for the candidate NSP is below bandwidth allocation for the candidate NSP;
• if the current bandwidth for the candidate NSP is below the bandwidth allocation for the candidate NSP, then
• comparing performance data for the candidate NSP with other ones of the remaining NSPs;
• and
• comparing cost data for the candidate NSP with other ones of the remaining NSPs;
• and
• based on the performance data and cost data comparisons, changing the data flow from the first NSP to one of the remaining NSPs.