Detecting and quantifying latency components in accessing cloud services
AssigneeMicrosoft Technology Licensing, LLC
InventorsSrinivasachakrapani Kotipalli, Brad Rutkowski, Paul James Andrew, Konstantin Ryvkin, Chittaranjan Sadanand Pattekar, Saurabh Vats
Life of this patent
Estimating the term…
Abstract
A latency processing system detects traffic at a cloud service end point and analyzes packets in the detected traffic to identify a network configuration of a client that is accessing the cloud service. Latency components corresponding to different parts of the network configuration are identified and quantified. A recommendation engine is controlled to generate and surface an output indicative of recommendations for reducing network latency.
Claims20
A computer system, comprising:
- at least one processor;
- and
- a data store storing instructions which, when executed by the at least one processor, cause the at least one processor to perform steps compromising:
- detecting a first latency for a first client, using a first set of network components, in connecting to a cloud service endpoint;
- detecting a second latency for a second client, using a second set of network components, in a common geographic area as the first client, in connecting to the cloud service endpoint using the second set of network components, the second latency being higher than the first latency;
- based on the first latency encountered using the first set of network components, identifying a change in the second set of network components that can be made by the second client to obtain a latency reduction for the second client;
- and
- generating a latency reduction recommendation output to the second client, the latency reduction recommendation output identifying:
- a quantity of the latency reduction available to the second client, and
- the identified change in the second set of network components.
A computer implemented method, comprising:
- detecting a first latency for a first client, using a first set of network components, in connecting to a cloud service endpoint;
- detecting a second latency for a second client, using a second set of network components, in a common geographic area as the first client, in connecting to the cloud service endpoint using the second set of network components, the second latency being higher than the first latency;
- based on the first latency encountered using the first set of network components, identifying a change in the second set of network components that can be made by the second client to obtain a latency reduction;
- and
- generating a latancy reduction recommendation output to the second client, the latency reduction recommendation output identifying;
- a quantity of the latency reduction available to the second client, and
- the identified change in the second set of network components.
A computer implemented method, comprising:
- detecting a plurality of different requests, originating from a common geographic area, at a cloud service endpoint;
- detecting a network configuration from which each request originated;
- detecting a latency encountered by each network configuration to identify a high latency network configuration having a first latency and a low latency network configuration having a second latency lower than the first latency;
- identifying a client that is part of the high latency network configuration;
- generating a network configuration recommendation for the identified client, the network configuration recommendation quantifying a reduction in latency available to the identified client by switching to the low latency network configuration;
- and
- controlling an interface generator to output the network configuration recommendation for surfacing to the identified client.
Dependencies are inferred from the claim text.