Amazon Kinesis Video Streams
Cloud service · AWS · Updated

Amazon Kinesis Video Streams

In short

Amazon Kinesis Video Streams is a fully managed AWS service that allows users to easily capture, process, and store video streams for analytics and machine learning applications.

Cloud Services Hub →

7 key features of Amazon Kinesis Video Streams

Amazon Kinesis Video Streams offers real-time video processing, serverless integration, secure storage, customizable playback, adaptive bitrate streaming, and global distribution for comprehensive video streaming and analytics.

01
Real-time Video Processing

Amazon Kinesis Video Streams enables users to securely stream video from connected devices to AWS for analytics, machine learning (ML), and other processing.

02
Serverless Integration

Easily integrates with serverless services and AWS ecosystems such as AWS Lambda, Amazon S3, and Amazon Rekognition for enhanced video analysis and storage solutions without managing infrastructure.

03
Secure and Durable Storage

Provides secure, durable, and scalable storage for streaming video, supporting encryption at rest and in transit to ensure highly sensitive video data is protected.

04
Customizable Playback

Supports video playback capabilities that can be customized and integrated into apps, allowing for the creation of time-based indexed archives for easy searching and seamless playback.

05
Global Distribution

Facilitates the creation of applications that require video streams to be available globally, reducing latency and improving the viewing experience for a worldwide audience.

06
Non-Video Data Ingestion

Supports non-video, time-serialized data such as audio, thermal imagery, depth data, and RADAR data.

07
Buffered and After-the-Fact Uploads

Producer libraries support streaming in real time, after buffering media for a few seconds, or through after-the-fact media uploads.

Where Amazon Kinesis Video Streams is used

Amazon Kinesis Video Streams is utilized for real-time surveillance, smart city analytics, telehealth, interactive live streaming, and machine learning model training, demonstrating its versatility across various domains.

  1. 01
    Real-Time Surveillance System

    Amazon Kinesis Video Streams enables the development of real-time surveillance systems where video streams from various locations can be securely ingested, processed, and played back for real-time monitoring, intrusion detection, and activity analytics. Users can easily store, access, and manage video feeds for surveillance across widely dispersed geographical locations.

  2. 02
    Live Video Analytics for Smart Cities

    This use case leverages Amazon Kinesis Video Streams for analyzing traffic patterns, public safety incidents, and urban planning processes in smart cities. By processing and analyzing live video feeds from city cameras, municipalities can make data-driven decisions to improve traffic flow, public safety, and emergency response times.

  3. 03
    Telehealth and Remote Patient Monitoring

    Amazon Kinesis Video Streams facilitates telehealth applications by enabling secure, live, and on-demand video interactions between healthcare providers and patients. This supports remote patient monitoring, virtual consultations, and physical therapy sessions, thus improving the accessibility and quality of healthcare services.

  4. 04
    Interactive Live Streaming

    For media and entertainment industries, Amazon Kinesis Video Streams supports the delivery of live events, gaming broadcasts, and interactive sessions in real-time. The platform offers low latency and high-quality video streams, enriching viewer experiences and enabling real-time interaction between content creators and their audiences.

  5. 05
    Machine Learning Model Training

    Amazon Kinesis Video Streams aids in the collection and preparation of diverse video feeds for machine learning model training. By providing a stream of real-world data, it accelerates the training and refinement of computer vision and other AI models, enabling more accurate predictions and functionalities in various applications such as autonomous driving and predictive maintenance.

How Amazon Kinesis Video Streams is priced

Amazon Kinesis Video Streams adopts a pay-as-you-go model for video ingest, storage, and consumption, with additional costs for optional features such as video analytics and data transfer.

Pay-as-you-go

Amazon Kinesis Video Streams pricing is based on the amount of data ingested, stored, and consumed through the service. Costs are incurred for video input (ingestion), video storage, and video output (consumption). This model allows for flexibility and scalability, ensuring you only pay for the resources you use.

Data Transfer

Standard AWS data transfer charges apply when data is retrieved from video streams to destinations outside AWS over the internet.

Storage Tier Ingestion

Hot-tier ingestion is charged per GB ingested, with no minimum retention period. Warm-tier ingestion is charged per 1,000 fragments persisted and has a minimum retention period of 30 days.

WebRTC Usage

WebRTC usage is billed for active signaling channels, signaling messages sent and received, TURN relay minutes, media ingested via WebRTC, and Multi-Viewer session viewing. TURN streaming, WebRTC ingestion, and Multi-Viewer viewing are metered in one-minute increments. WebRTC ingestion is charged per minute instead of the per-GB ingestion rate.

Image Generation

Images generated from streams are charged per million images, with separate pricing for resolutions of 1080p or lower and resolutions greater than 1080p.

Services Amazon Kinesis Video Streams works with

Amazon CloudWatchAmazon CloudWatch

Amazon Kinesis Video Streams publishes metrics to CloudWatch for ingestion, playback, storage, and API operations. You can create alarms to detect ingestion failures, increased latency, or a camera that has stopped sending video.

Amazon CognitoAmazon Cognito

Amazon Kinesis Video Streams applications can use Cognito identity pools to obtain temporary AWS credentials for authorized viewers and producers. The Kinesis Video Streams WebRTC JavaScript sample supports this authentication approach for accessing signaling channels.

Amazon ConnectAmazon Connect

Amazon Connect live media streaming sends customer call audio to Kinesis Video Streams. Applications can consume the audio during a call for processing such as transcription or real-time analysis.

Amazon EventBridgeAmazon EventBridge

Amazon Kinesis Video Streams API activity recorded by CloudTrail can be matched by EventBridge rules. For example, a rule can start an automation workflow when a video stream is created or deleted.

Amazon RekognitionAmazon Rekognition

Amazon Kinesis Video Streams supplies live video to Amazon Rekognition stream processors for supported streaming video analysis. Rekognition can analyze camera feeds for use cases such as recognizing faces or detecting people, pets, and packages.

Amazon S3Amazon S3

Amazon Kinesis Video Streams can automatically extract images from video and deliver them to an S3 bucket through an image generation configuration. You can configure the sampling interval, image format, and destination prefix.

Amazon SageMakerAmazon SageMaker

Amazon Kinesis Video Streams provides integration tooling that reads video fragments, extracts frames, and submits them to SageMaker inference endpoints. This lets custom machine learning models analyze live camera feeds.

Amazon Simple Notification Service (Amazon SNS)Amazon Simple Notification Service (Amazon SNS)

Amazon Kinesis Video Streams can publish fragment notifications to an SNS topic when media fragments are persisted. Subscribers can use those notifications to begin processing newly available video.

AWS CloudFormationAWS CloudFormation

Amazon Kinesis Video Streams video streams and WebRTC signaling channels can be provisioned as CloudFormation resources. Templates can define stream retention, encryption settings, and tags alongside the application's other infrastructure.

AWS CloudTrailAWS CloudTrail

Amazon Kinesis Video Streams integrates with CloudTrail to record supported API activity, including control-plane operations on streams and signaling channels. These records help audit who created, changed, or deleted resources.

Amazon Kinesis Video Streams exam questions

Real practice exam questions on Amazon Kinesis Video Streams, with the correct answer and its explanation. Answer them with every option in the exam simulator.

MLA-C02 (beta)

A company collects video from inspection cameras to prepare ML training datasets. Each edge gateway buffers encoded video fragments on local disk and uploads them to AWS. The cloud ingestion layer must retain the video for 24 hours. Network connections sometimes fail during uploads, so a gateway must not delete a local fragment until the cloud service confirms that the fragment has been durably stored. The company wants to use managed video ingestion without building a separate storage-confirmation system.

Which implementation meets these requirements?

Show answerCORRECT ANSWER

Use the Amazon Kinesis Video Streams Producer SDK with 24-hour retention. Delete each local fragment after receiving its PERSISTED fragment acknowledgment.

With retention enabled, a PERSISTED acknowledgment confirms that the corresponding fragment has been durably stored. Tracking this acknowledgment for each fragment allows the gateway to safely remove its local copy without implementing a separate storage-confirmation system.

SAA-C03

A company operates hundreds of cameras at remote construction sites. The cameras produce continuously encoded H.264 video. Site supervisors need browser-based viewing of live feeds and must be able to select a camera and timestamp to review footage from the previous 48 hours. A separate analytics application must also be able to retrieve the recorded video for processing. The company wants to avoid maintaining custom video-fragment indexing and playback-packaging infrastructure.

Which solution meets these requirements with the LEAST operational overhead?

Show answerCORRECT ANSWER

Use the producer SDK to ingest each camera's video into Amazon Kinesis Video Streams with 48-hour retention. Use its HLS playback APIs for live and historical viewing and archived-media APIs for analytics retrieval.

Kinesis Video Streams manages video ingestion, retention, and time-based indexing. Its HLS playback APIs support live and historical browser playback, and its archived-media APIs allow applications to retrieve retained footage without building a custom video storage and indexing system.

Practice Amazon Kinesis Video Streams in a full exam

Questions on Amazon Kinesis Video Streams in each practice exam, with every answer option and an explanation for each:

Start a free practice exam

What is Amazon Kinesis Video Streams

3 min read

Amazon Kinesis Video Streams is a highly scalable and secure service provided by AWS that allows customers to stream live video from devices to the AWS Cloud for analytics, machine learning (ML), playback, and other processing. It is designed to make it easy to securely and reliably stream video from connected devices to AWS without requiring infrastructure or hardware investments, enabling applications across a wide range of industries including home security, automation, industrial monitoring, and video content delivery. 

The service simplifies the task of consuming and processing video streams at any scale, offering real-time and batch capabilities. It efficiently handles the ingestion, storage, and retrieval of video streams, allowing developers to build applications that can extract meaningful insights from the video content using other AWS services like Amazon Rekognition for video analysis or integrate with Amazon SageMaker for custom ML model inference. 

Multiple independent applications can consume the same Kinesis video stream in parallel.

Kinesis Video Streams offers two storage tiers: hot storage for frequent access, live playback, and real-time analytics, and warm storage for cost-effective extended media retention with sub-second access latency.

This seamless integration enables a myriad of scenarios such as facial recognition, movement detection, or any other analytical operation that can be applied to video data. Amazon Kinesis Video Streams ensures the privacy and security of video streams through industry-standard encryption methods. It encrypts data at rest and in transit, ensuring end-to-end privacy and security for its users. This is crucial for applications handling sensitive video material, ensuring compliance with relevant regulations and maintaining customer trust. 

One of the key benefits of using Amazon Kinesis Video Streams is its adaptability to various video formats and codecs, which ensures broad device compatibility ranging from simple webcams to sophisticated camera systems equipped with the latest technology. The service's ability to process streams from millions of devices simultaneously makes it an ideal platform for applications like city-wide surveillance systems, live event broadcasting, and extensive industrial monitoring setups. 

Moreover, its integrated analytics and machine learning features not only allow for real-time video content analysis but also simplify the development process. Developers can focus on building their applications without worrying about the complexities of video streaming infrastructure. 

Amazon Kinesis Video Streams provides SDKs and APIs that facilitate the integration of video streaming capabilities into any application, enabling a smoother development experience. With its robust feature set, Amazon Kinesis Video Streams represents a powerful tool for businesses and developers looking to leverage video content in real time or through batch processing. 

Traffic through a VPC endpoint has a separate request-rate quota of 500 requests per second per AWS account per endpoint, shared across ingestion and playback APIs. Requests above this limit return ClientLimitExceededException; clients should use exponential backoff and retry. All other Kinesis Video Streams quotas also apply to endpoint traffic.

Kinesis Video Streams also supports WebRTC streaming of real-time audio and video to up to three concurrent viewers while recording the session to the cloud for storage, playback, and analytical processing. Participants can engage in audio conversations with each other.

Whether it's for security purposes, consumer applications, or industrial monitoring, Amazon Kinesis Video Streams delivers a flexible, secure, and efficient solution for streaming video content to the AWS cloud, unlocking a wealth of possibilities for analyzing and processing video data at scale.

Learn Amazon Kinesis Video Streams in lessons

3 short lessons on Amazon Kinesis Video Streams in 2 certification courses.

Open the lessons

AWS architecture icons are the property of Amazon Web Services. BlowStack is not affiliated with or endorsed by AWS.