Abstract: A polytope is generated, based on expert input, in an output parameter space. The polytope constrains network parameters to value ranges that are a subset of possible values represented in the output parameter space. Network traffic data associated with data requests to computer applications based on static policies is collected over a time block. Each static policy in the plurality of static policies comprises parameter values, for network parameters in the set of network parameters, that are constrained to be within the polytope. Machine learning is used to estimate best parameter values for the network parameters that are constrained to be within the polytope. The best parameter values are verified by comparing to parameter values determined from a black box optimization. The best parameter values are propagated to be used by user devices to make new data requests to the computer applications.

Abstract: A data-driven approach to network performance diagnosis and root-cause analysis is presented. By collecting and aggregating data attribute values across multiple components of a content delivery system and comparing against baselines for points of inspection, network performance diagnosis and root-cause analysis may be prioritized based on impact on content delivery. Recommended courses of action may be determined and provided based on the tracked network performance analysis at diagnosis points.

Abstract: A polytope is generated, based on expert input, in an output parameter space. The polytope constrains network parameters to value ranges that are a subset of possible values represented in the output parameter space. Network traffic data associated with data requests to computer applications based on static policies is collected over a time block. Each static policy in the plurality of static policies comprises parameter values, for network parameters in the set of network parameters, that are constrained to be within the polytope. Machine learning is used to estimate best parameter values for the network parameters that are constrained to be within the polytope. The best parameter values are verified by comparing to parameter values determined from a black box optimization. The best parameter values are propagated to be used by user devices to make new data requests to the computer applications.

Abstract: Network traffic data associated with computer applications is collected based on static policies. First network parameter vectors are generated over a time period. Each network parameter vector of the first network parameter vectors comprises first optimal values, estimated by a Bayesian learning module using a generative model, for network parameters. Second network parameter vectors are generated over the same time period. Each network parameter vector of the second network parameter vectors comprises second optimal values, computed by a best parameter generator through optimizing an objective function, for the network parameters. It is determined whether the first network parameter vectors converge to the second network parameter vectors and whether network parameter optimization for the network parameters is performing normally.

Abstract: Network traffic data associated with computer applications is collected based on static policies. First network parameter vectors are generated over a time period. Each network parameter vector of the first network parameter vectors comprises first optimal values, estimated by a Bayesian learning module using a generative model, for network parameters. Second network parameter vectors are generated over the same time period. Each network parameter vector of the second network parameter vectors comprises second optimal values, computed by a best parameter generator through optimizing an objective function, for the network parameters. It is determined whether the first network parameter vectors converge to the second network parameter vectors and whether network parameter optimization for the network parameters is performing normally.

Abstract: Data analysis is performed through a series of commands that apply functions to an initial scope of data. In a client-server architecture, a data analyst may interact with and view a scope of data through a series of commands. Query formation may be performed at a server to generate reports of data to be presented at the client.

Abstract: A dynamic approach to optimizing configuration of network parameters is presented. By gathering operational contexts and aggregating optimized network performance data against a baseline, a training data set may be generated. Client-side policies are determined, in part, by applying machine learning techniques on the training data set to achieve desired outcomes. Data delivery strategies are compiled at user devices to deliver content using the optimized network configuration values based on the operating contexts.

Abstract: An adaptive multi-phase approach to estimating network parameters is presented. By gathering and aggregating raw network traffic data and comparing against default network parameters, a training data set may be generated. A black box optimization may be used in tandem with a supervised learning algorithm to bias towards better choices and eventually pick network parameters which optimize performance. Data delivery strategies are applied to deliver content using the optimized network policies based on the estimated parameters.

Abstract: Network traffic data associated with computer applications is collected based on static policies. First network parameter vectors are generated over a time period. Each network parameter vector of the first network parameter vectors comprises first optimal values, estimated by a Bayesian learning module using a generative model, for network parameters. Second network parameter vectors are generated over the same time period. Each network parameter vector of the second network parameter vectors comprises second optimal values, computed by a best parameter generator through optimizing an objective function, for the network parameters. It is determined whether the first network parameter vectors converge to the second network parameter vectors and whether network parameter optimization for the network parameters is performing normally.

Abstract: Network traffic data associated with data requests to computer applications is collected. Specific values for specific scope-level fields are used to identify a specific scope. Traffic shares for combinations of values for specific sub-scope-level fields are determined. Based on the traffic shares, specific sub scopes are identified within the specific scope. It is determined whether customized network strategies developed specifically for the specific sub scopes are to be applied to handling new data requests that share the specific values for the specific scope-level fields and the specific combinations of values for the specific sub-scope-level fields. In response to determining that a customized network strategy for a sub scope is to be applied, estimated optimal values for network parameters in the customized network strategy are to be used by user devices to make new data requests to the computer applications.

Abstract: Network traffic data associated with data requests to computer applications based on static policies is collected. An optimization order is established among network parameters. A first network parameter of a higher rank in the optimization order is estimated based on the collected network traffic data before one or more other network parameters of lower ranks are estimated. Optimal values for the other network parameters are estimated based at least in part on the estimated first optimal value for the first network parameter. The estimated first optimal value of the first network parameter and the estimated optimal values for the other network parameters are propagated to be used by user devices to make new data requests to the computer applications.

Abstract: An data driven approach to generating synthetic data matrices is presented. By retrieving historical network traffic data, probabilistic models are generated. Optimal distribution families for a set of independent data segments are determined. Applications are tested and performance metrics are determined based on the generated synthetic data matrices.

Abstract: A polytope is generated, based on expert input, in an output parameter space. The polytope constrains network parameters to value ranges that are a subset of possible values represented in the output parameter space. Network traffic data associated with data requests to computer applications based on static policies is collected over a time block. Each static policy in the plurality of static policies comprises parameter values, for network parameters in the set of network parameters, that are constrained to be within the polytope. Machine learning is used to estimate best parameter values for the network parameters that are constrained to be within the polytope. The best parameter values are verified by comparing to parameter values determined from a black box optimization. The best parameter values are propagated to be used by user devices to make new data requests to the computer applications.

Abstract: An data driven approach to emulating application performance is presented. By retrieving historical network traffic data, probabilistic models are generated to simulate wireless networks. Optimal distribution families for network values are determined. Performance data is captured from applications operating on simulated user devices operating on a virtual machine with a network simulator running sampled tuple values.

Abstract: Using the ALTO Service, networking applications can request through the ALTO protocol information about the underlying network topology from the ISP or Content Provider. The ALTO Service provides information such as network resource preferences with the goal of modifying network resource consumption patterns while maintaining or improving application performance. This document describes, in one example, an ALTO server that implements enhancements to the ALTO service to assign a PID-type attribute to each of a set of one or more PIDs each associated with a subset of one or more endpoints of a network, wherein a PID-type attribute specifies a type for the subset of endpoints associated with the PID. The ALTO server generates an ALTO network map that includes a PID entry to describe each of the PIDs, wherein each PID entry includes a PID-type field that stores the assigned PID-type attribute for the PID described by the PID entry.

Abstract: A data-driven approach to network performance diagnosis and root-cause analysis is presented. By collecting and aggregating data attribute values across multiple components of a content delivery system and comparing against baselines for points of inspection, network performance diagnosis and root-cause analysis may be prioritized based on impact on content delivery. Recommended courses of action may be determined and provided based on the tracked network performance analysis at diagnosis points.

Abstract: Network performance data metrics are gathered and aggregated. A policy engine chooses an optimal selection of networking protocol based on the metrics. Data delivery strategies are applied to a portion of a network to deliver content using the received choice of networking protocol policy optimized by machine learning techniques.

Abstract: A data-driven approach to network performance diagnosis and root-cause analysis is presented. By collecting and aggregating data attribute values across multiple components of a content delivery system and comparing against baselines for points of inspection, network performance diagnosis and root-cause analysis may be prioritized based on impact on content delivery. Alerts may be generated to present recommended courses of action based on the tracked performance analysis.

Abstract: An adaptive multi-phase approach to estimating network parameters is presented. By gathering and aggregating raw network traffic data and comparing against default network parameters, a training data set may be generated. A black box optimization may be used in tandem with a supervised learning algorithm to bias towards better choices and eventually pick network parameters which optimize performance. Data delivery strategies are applied to deliver content using the optimized network policies based on the estimated parameters.

Abstract: In general, techniques are described for dynamically generating attributes from routing topology information and assigning dynamically generated attributes to network map entries to further characterize PIDs described therein. For example, a provider or other entity assigns, within a network device, endpoint types to one or more address prefixes for which the network device originates or forwards route advertisements. For each typed prefix, the network device adds an endpoint type identifier for the assigned endpoint type to route advertisements that traverse or originate with the network device and specify the prefix. An ALTO server peers with router advertisers to receive route advertisements. When the ALTO server receives a route advertisement that includes an endpoint type identifier, the ALTO server maps the endpoint type identifier to a PID attribute and assigns the PID attribute to a PID that includes a prefix identified in the route advertisement.