Amazon AppStream 2.0 is an AWS End User Computing (EUC) service optimized for software-as-a-service (SaaS) application streaming, converting desktop applications to SaaS, and virtual desktop use cases. AppStream 2.0 works with Amazon WorkSpaces Thin Client—improving end-user and IT staff productivity with cost-effective, easy-to-manage, secure access to AWS End User Computing services.

AppStream 2.0 lets you stream SaaS applications or convert desktop apps to SaaS without rewriting them. When used to stream applications, it’s easy to install your applications on AppStream 2.0, set launch configurations, and make your applications available to users. AppStream 2.0 offers a wide selection of configuration options so you can select the instance type and auto-scaling parameters that best match your application and end-user requirements. AppStream 2.0 allows you to launch applications in your own network, which means your applications can interact with your existing AWS resources.

AppStream 2.0 can also be configured to securely deliver virtual desktops with selective persistence. Selective persistence means IT curates the user experience by controlling applications, application settings, operating system, upgrades, and patches deployed via a golden image. Every time the user logs on, they receive a fresh virtual desktop based on the golden image. With selective persistence, the user saves their files in a central location (home folder), while application settings (browser favorites, webpage sessions, application connection profiles, plugins, UI customizations) and storage connections (Google Drive, OneDrive for Business, Amazon S3) remain persistent between user sessions. No data is retained on the virtual desktop between user sessions. See additional detail below on persistent storage and persistent settings.

AppStream 2.0 gives the IT admin flexibility to provision one or more user session per compute instance. AppStream 2.0 multi-session capability allows multiple users to share compute, memory, storage, and system software associated with a given instance, while continuing to enjoy the ability to auto-scale resources based on actual usage. This helps reduce solution cost and maximize resource utilization. Multi-session capability gives IT administrators the ability to adjust the number of user sessions per instance based on ther equirements of the individual users—allowing support for a wide range of user types without over provisioning. For multi-session fleets, IT admins can use out-of-the-box AppStream2.0 / Amazon CloudWatch consoles to monitor performance of user sessions and utilization of underlying resources. In an On-Demand fleet, AppStream 2.0 multi-session capability reduces the average time it takes for users to set up their sessions on a given multi-session instance.

Multi-session functionality is supported only for Windows configurations of Always-On and On-Demand fleets. Multi-session instances are charged on an hourly basis, regardless of the number user sessions running per instance. See pricing page for more detail.

Simple application management

Amazon AppStream 2.0 enables you to quickly and easily install, test, and update your applications using the image builder. Any application that runs on Microsoft Windows Server 2012 R2, Windows Server 2016, Windows Server 2019, or Amazon Linux 2 is supported and you don’t need to make any modifications. When your testing is complete, you can set application launch configurations, configure default user settings, and publish your image for users to access. Using managed AppStream 2.0 image updates, you can keep your images up to date with the latest operating system updates, driver updates, and AppStream 2.0 agent software.

Multiple instance types

Amazon AppStream 2.0 runs your applications on streaming instances, which provide the GPU, CPU, memory, storage, and networking capacity you need. With AppStream 2.0, you can choose from a selection of them to optimize the performance of your applications, and reduce your overall costs. You can use General Purpose, Compute Optimized, Memory Optimized, Graphics Design, Graphics Pro, and Graphics G4 families based on your use case. Users access all their applications from the same streaming instance, and a streaming instance is assigned to only one user at any time.

GPU optimized instances

Amazon AppStream 2.0 offers Graphics Design, Graphics Pro and Graphics G4 and Graphics G5 instance families.

Graphics Design instances are ideal for delivering applications such as Adobe Premiere Pro, Autodesk Revit, and Siemens NX that rely on hardware acceleration of DirectX, OpenGL, or OpenCL. Powered by AMD FirePro S7150x2 Server GPUs and equipped with AMD Multiuser GPU technology, instances start from 2 vCPU, 7.5 GiB system memory, and 1 GiB graphics memory, to 16 vCPUs, 61 GiB system memory, and 8 GiB graphics memory.

Graphics G5 instances are based on the EC2 G5 family. Amazon EC2 G5 instances feature NVIDIA A10G Tensor Core GPUs and 2nd generation AMD EPYC processors, delivering up to 3x higher performance for graphics-intensive applications and machine learning inference, and up to 3.3x higher performance for training simple to moderately complex machine learning models compared to Amazon EC2 G4dn instances. AppStream 2.0 offers seven different instance sizes for Graphics G5, ranging from xlarge with 4 vCPUs and 16 GB of system memory to 24xlarge with 96 vCPUs and 384 GB of system memory.

Graphics g4dn instances are based on the EC2 G4 family. Amazon EC2 g4dn instances deliver the industry’s most cost-effective and versatile GPU instance for running graphics-intensive applications on AWS. G4dn instances provide the latest generation NVIDIA T4 GPUs, AWS custom Intel Cascade Lake CUs, up to 100 Gbps of networking throughput, and up to 1.8 TB of local NVMe storage. These instances are ideal for streaming graphics intensive applications that rely on NVIDIA GPU libraries such as CUDA using AppStream 2.0. AppStream 2.0 offers six different g4dn instance sizes, ranging from 4 vCPUs and 16 GiB of memory to 64 vCPUs and 256 GiB of memory

The Graphics Pro instance family offers three different instance types to support the most demanding graphics applications. Powered by NVIDIA Tesla M60 GPUs with 2048 parallel processing cores, there are three Graphics Pro instances types starting from 16 vCPUs, 122 GiB system memory, and 8 GiB graphics memory, to 64 vCPUs, 488 GiB system memory, and 32 GiB graphics memory. These instance types are ideal for graphic workloads that need a massive amount of parallel processing power for 3D rendering, visualization, and video encoding, including applications such as Petrel from Schlumberger Software, Landmark's DecisionSpace, or MotionDSP's Ikena. For more information on available instance types and pricing, see Amazon AppStream 2.0 pricing.

Amazon VPC support

With Amazon AppStream 2.0, your desktop applications can be launched inside an Amazon Virtual Private Cloud (VPC). You can use VPC security groups to provide granular access control to streaming instances, and to manage users' access to the resources in your VPC, such as your databases, file shares, license servers, and application servers. You can also specify how your desktop applications access the public internet, just as you would with any other resource in your VPC.

Identity federation

Amazon AppStream 2.0 supports federated sign-in using SAML 2.0. Users can sign in to AppStream 2.0 using their existing credentials, and start streaming applications. As an administrator, you can use your existing user directory to control end-user access to applications available via AppStream 2.0. You can quickly add or remove access for users or groups, restrict access based on user locations, and enable multi-factor authentication. You can enable federated access and controls via any SAML 2.0 compliant identity provider. For a passwordless login experience, Amazon AppStream 2.0 supports certificate-based authentication for deployments that are joined to your Active Directory.

Microsoft Active Directory domain support

Your Amazon AppStream 2.0 Always-On and On-Demand fleet streaming instances and image builders can connect to your Microsoft Active Directory (AD) domain. This allows you to apply your existing AD policies to your streaming instances, and provides your users with single sign on access to Intranet sites, file shares, and network printers from within their applications. Your users are authenticated using a SAML 2.0 provider of your choice, and can access applications that require a connection to your AD domain.

Smart card support

Your users can use their Personal Identity Verification (PIV) and Common Access Card (CAC) smart card, and other types of smart cards, to sign in to a Windows OS based AppStream 2.0 streaming instance that is joined to a Microsoft Active Directory domain. Smart cards are also supported for in-session authentication for streaming applications. When this feature is enabled and your user streams using the Windows native client, AppStream 2.0 redirects smart card requests from the streaming instance to a smart card reader that is connected to the user’s local computer. The smart card and smart card reader remain accessible to both the streaming instance and the user’s local computer.


Amazon AppStream 2.0 offers a variety of fleet options, allowing you to configure the optimal solution for your specifc requirements. The table below provides a high-level overview of each available fleet type.

Fleet type Description Use case Pricing
Always on - Instant connection
- Standby instances are running
- Optimized for instant-on experience 
When instant startup is required
- Subscription applications
- Trials and demos 
- Windows: per hour
- Linux: per second 
On demand - Applications launch ~2 min
- Standby instances powered off
- Optimized for cost
When cost savings are critical and usage is predictable
- Classroom and lab
- Contact center 
- Windows: per hour
- Linux: per second
- Small hourly fee for powered off instances 

- AWS managed pool streaming instances
- No standby capacity to pay for
- No need to manage AMI 

When utilization is unpredictable
- Trainings
- Game days 
- Windows: per second
- Linux: per second 

Elastic fleets

Elastic fleets are a serverless fleet type that allows you to deliver your applications to end users without needing to predict concurrency, create or manage auto scaling policies, or create any images. Your applications are stored within virtual hard disks that are downloaded to streaming instances on user request simplifying how you deliver streaming applications to users. Elastic fleets allow you to focus on delivering your application rather than managing fleet capacity and are best suited for delivering portable applications and when usage patterns may not be constant, like delivering trials, trainings, and demos, or converting your traditionally delivered desktop application to a software as a service.

AppStream 2.0 includes applications manager. Applications manager is designed to help you package and mobilize Windows applications across environments, without affecting the underlying base image or operating system. After launching an app block builder instance, simply install your application(s) and create a self-contained and sharable application image. This application image is sharable across multiple AppStream 2.0 elastic fleets provisioned across different AWS Regions

Fleet auto scaling

With Always-On and On-Demand fleets, you can use auto scaling policies to adjust the number of instances that are running to reduce your streaming costs. Each new Always-On and On-Demand fleet you create is configured default scale out and scale in policies, which you can customize to meet your user requirements. You can also create your own scaling policies for each of your Always-On and On-Demand fleets. You can use schedule-based, usage-based, or both types of policies to optimize the number of instances your fleet has.

With Elastic fleets, AppStream 2.0 manages the size of the fleet for you without needing to use autoscaling policies.


Amazon AppStream 2.0 allows you to monitor the utilization of your AppStream 2.0 fleet resources using Amazon CloudWatch metrics. With Elastic fleets, you can track the number of instances that are being used over time. With Always-On and On-Demand fleets, you can see the size of your fleet, the number of instances you have running, and the available capacity for new connections. You can also track utilization over time, which helps you ensure that your fleet is the correct size. Using CloudWatch, you can set alarms to notify you when there is insufficient capacity to support your users, and to react to changes in your Always-On and On-Demand fleet.

Programmatic control

Amazon AppStream 2.0 includes APIs that you can use to easily integrate and extend the service. The APIs enable you to create, update, and delete AppStream 2.0 resources, and provide detailed information about resource states. You can create URLs for administrators to connect to their image builders to install applications, and create URLs for users to access their AppStream 2.0 applications.

Browser and client access

Amazon AppStream 2.0 allows you to access your desktop applications from HTML5-capable browsers such as Google Chrome, Mozilla Firefox, Microsoft Internet Explorer, and Microsoft Edge. No plug-ins are needed. You can also access your desktop applications using the AppStream 2.0 Windows Client to use dual monitors and USB peripherals such as 3D mice. The Windows Client also supports keyboard shortcuts, such as Alt + Tab, clipboard shortcuts, and function keys. The Windows Client and Chromium-based web browsers support real-time audio-video (AV) when connecting to AppStream 2.0 sessions.

Persistent storage

Amazon AppStream 2.0 offers multiple options for persistent file storage to allow users to store and retrieve files between their application streaming sessions. You can use a home folder backed by Amazon S3, Google Drive for G Suite, or Microsoft OneDrive for Business. Each of these are accessed from the my files tab within an active AppStream 2.0 streaming session, and content can be saved or opened directly from the File menu in most apps.

Persistent settings

You can enable persistent application and Windows settings for your users on AppStream 2.0. Your users' plugins, toolbar settings, browser favorites, application connection profiles, and other settings will be saved and applied each time they start a streaming session. Your users' settings are stored in an S3 bucket you control in your AWS account.

Simple user interface

Amazon AppStream 2.0 offers an intuitive user interface, making it easy to control your experience. You can use the application-centric view, or use the standard desktop experience. You can run multiple applications and easily switch and interact with them. You can copy and paste between applications and your local machine, quickly upload or download files, and print to your local or network printer. You can also listen to audio output from applications, and control the bandwidth and fidelity for the experience. With the AppStream 2.0 Windows Client, your applications integrate with your local machine, providing a seamless experience between locally installed applications and remote applications from AppStream 2.0.

NICE DCV protocol

Amazon AppStream 2.0 uses NICE DCV technology to provide secure, high-performance access to your applications. NICE DCV delivers a fluid interactive experience for business, design, engineering, and software development applications. It delivers a secure pixel stream to your end users and always adjusts to network conditions to provide an optimal end-user experience. NICE DCV secures both pixels and end-user inputs using end-to-end AES-256 encryption.

HTTPS access

With Amazon AppStream 2.0, your application streams and user input flows through a secure streaming gateway on AWS over HTTPS. Streaming instances are not directly accessible from the internet, and users can only access their applications through the streaming gateway after being authenticated.

Globally available

Amazon AppStream 2.0 is available in multiple AWS regions globally. You can host your AppStream 2.0 resources in multiple AWS regions, and direct users to the closest AWS region for the best end-user experience.

Learn more about Amazon AppStream 2.0 pricing

Visit the pricing page
Ready to start?
Get started with Amazon AppStream 2.0
Have more questions?
Contact us