Amazon Technologies, Inc.

Stefan Christian Richter

Marc Joliveau

Nicholas James Benson

Kenneth Thomas Burford

Amarsingh Buckthasingh Winston

Li Tong

Yongjun Wu

Justin Michael Binns

Lei Li

James Marvin Freeman, Ii

Euan McLeod

Charles Benjamin Franklin Waggoner

Orlando Maldonado

Sandeep Bajaj

Andrew Stuart Glass

Jesse Jerome Rosenzweig

Gregory Scott Benjamin

Michael Reddick Coleman

Srikanth Kiran Kotagiri

Kyle Bradley Koceski

Dustin Encelewski

Chris Longo

Rakesh Tangirala

Paul Nann

David Alonso Flamini

Ali Khayam

Bobby Brown

Lincoln Travis Dale

Tahir Azim

Andrey Sokolov

Benjamin Villain

Apoorv Srivastava

Michael Robert Starr

Elliot Fanwar George

Ryan Dell Olson

A system is disclosed for programming target devices with a filtered (reduced) set of route updates. The system can be divided into a backend system and a frontend system. The backend system analyzes all routes being used and generates a prefix list used for filtering. The frontend system can filter the route updates to target routers and switches to ensure that the number of prefixes installed in a forwarding table of the target routers or switches does not exceed a configured amount. Different frontend systems can coordinate to prevent deviations in the forwarding tables. More specifically, different frontend systems can synchronize with each other to ensure that a union of hot prefixes they propagate always remains under a configured prefix count. The backend system can be responsible for computing and maintaining a sorted list of prefixes, which can be at a per-site or per-service granularity.

Network devices programmed with most-used routes

Programming routers in a network is accomplished using a global network manager and a plurality of local network managers. The global network manager analyzes an entire network and demand matrix in order to provide restrictions and recommendations that are communicated to a set of the local network managers. The local network managers can use the restrictions and recommendations to make local routing decisions defining a percentage of traffic routed to each neighboring node. The local network manager collects traffic metrics, such as local topology and traffic information by measuring the capacity available to reach any neighboring node. The local network manager uses the metrics in combination with the restrictions and recommendations from the global manager to make decisions on how to route network traffic.

Directing network traffic using local routing decisions with a global overview

A backbone service exposes network parameters, such as a minimum available bandwidth, latency, or packet loss, in a tunnel path between any source-destination pairs. The network parameters can be mapped as a function of time so that service teams can schedule when to use the backbone with minimized interruption to other users. The data generated by the backbone service can be transmitted, stored or displayed for informational purposes to provide insights to service teams on how to better leverage the network and create awareness of the current status of the backbone. The backbone service can be extended to provide bandwidth brokerage for controlling traffic distribution in the network. The backbone service can further provide triggered messages that inform service teams about failures in the network that could reduce the available bandwidth. The messages can further target users of affected source-destination pairs.

Backbone network performance analysis

Systems and methods are disclosed for generating synthetic representations of a network and performing path computations using the synthetic representations. A model of the network is created including different representations for different regions of the network based on the network mesh patterns (e.g., a sparse mesh representation, a Clos mesh representation, and a flat mesh representation). The generated representations include synthetic, aggregated nodes and/or links in the represented region that are determined based on different processes according to the type of mesh in the region. Path computations are performed for each representation (e.g., in parallel), then joined to form end-to-end paths between a source and a destination. The computed paths may be used to select a path for routing data through the network.

Computation of ranked path options in networks

Techniques are described for adjusting parameter settings for bitrate selection algorithms for different segments of a population of devices streaming content. Streaming sessions are identified according to session characteristics. Within each segment of sessions, control parameter settings are sent to devices corresponding to a subset of each segment. Test parameter settings are sent to devices corresponding to another subset of each segment. If the test parameter settings result in better playback performance relative to the control parameter settings, the test parameter settings become the new control parameter settings, and new test parameter settings are generated.

Adjusting parameter settings for bitrate selection algorithms

A system for delivering live streaming content based on accurate media data fragment size and duration. A media server may generate custom data identifying a size and duration of a current fragment of a media file. The media server may insert the custom data (e.g., as a custom header or unique packet identifier) and send the custom data to a client media player. The client media player may be configured to decode the custom data and determine the current fragment size and duration.

Streaming media file management

Techniques are described for server-side fragment insertion and delivery of media content. After receiving a request to play back media content, a manifest is generated that includes fragments associated with the requested media content and fragments associated with different media content. A client device requests fragments from a media server according to the manifest, and the media server provides the correct fragment in response to the request.

Server-side fragment insertion and delivery

Techniques are described for switching streaming of media content between content delivery networks (CDNs) during mid-stream. An initial CDN can be used to stream the media content. The performance of a second CDN can be measured by requesting a fragment of the media content as a measurement or test fragment. Streaming of the media content can be switched to the second CDN if its performance is determined to be better than the initial CDN.

Mid-stream content delivery network switching

Techniques are described for reducing the delay between the live playhead of live streaming content and the client playhead of a client device consuming the live stream. In one technique, an increased playback speed is used by the media player on the client device so that the delay is gradually reduced. In another technique, the media player jumps forward in the stream, skipping content identified as expendable.

Catching up to the live playhead in live streaming

Techniques are described that enable the use of variable bit rate (VBR) encoding for live content.

VBR encoding of live content

Techniques are described for adjusting parameter settings for bitrate selection algorithms for devices streaming media content. Control parameter settings are selected for playback of first media content. Test parameter settings are selected for playback of second media content. If the test parameter settings result in better playback performance relative to the control parameter settings, the test parameter settings become the new control parameter settings.

Techniques are described for balancing traffic among content delivery networks (CDNs). A CDN balancer can analyze CDN performance data and determine CDN balancing ratios used to distribute traffic among the CDNs.

Content delivery network balancer

Techniques for delivering live streaming content are described by which a streaming heuristic is selected based on the delay between the live and client playheads. For live content for which the delay is short, a low latency streaming heuristic is selected which is designed to handle live content streams for which only a few seconds of content can be buffered. For live content for which the delay is longer, a higher latency streaming heuristic is selected which is designed to handle streams for which the delay is sufficient to build a buffer of content fragments.

Heuristics for streaming live content

A system for delivering live streaming content based on accurate media data fragment size and duration. The system may include a client media player to receive a portion of a streaming media file (e.g., in an MP4 format), download a first sub-portion of the streaming media file including fragment-level metadata, and parse and analyze the fragment-level metadata to determine a size and duration of a current fragment of the media file. A media server may generate custom data identifying a size and duration of a current fragment of a media file. The media server may insert the custom data (e.g., as a custom header or unique packet identifier) and send the custom data to a client media player. The client media player may be configured to decode the custom data and determine the current fragment size and duration.

Techniques are described for live streaming media content using on-demand manifests. Many manifest files providing playback options for different time periods of a live stream of media content can be generated and provided to viewer devices. Requests from viewer devices using the manifest files can be analyzed to determine a fragment to be provided for playback in relation to the time period of the live stream to be provided for playback.

Growing community of inventors

Seattle, WA, United States of America

Stefan Christian Richter