Desktop and Application Streaming
Amazon AppStream 2.0 releases a simple pricing tool
This blog walks you through how to use the AppStream 2.0 Simple Pricing Tool for Always-On and On-Demand fleets with common pricing examples. To learn about Elastic fleets, read AWS announces Availability of Elastic fleets with Amazon AppStream 2.0 and refer to the AppStream 2.0 pricing page.
Amazon AppStream 2.0 is a fully managed application streaming service that enables you to centrally manage your desktop applications and securely deliver them to any computer. With AppStream 2.0, you can easily scale to any number of users across the globe without acquiring, provisioning, and operating hardware or infrastructure. Software vendors are using AppStream 2.0 to stream GPU-intensive 3D design and engineering applications to users, while enterprises are replacing their application streaming technologies with AppStream 2.0.
Customers really like the AppStream 2.0 pricing model of paying only for the instances they provisioned to meet their users’ streaming needs, but it is different than their existing application streaming environments that are based on provisioning for peak capacity, even during nights, weekends, and holidays when load would be lower. To help customers estimate their AppStream 2.0 price, we developed a simple, easy to use pricing tool. Simply provide your total number of users, actual concurrent usage per hour, instance type, and fleet utilization, and the pricing tool will estimate your per user price. It will also show you the estimated price savings when you use an On-Demand fleet instead of an Always-On fleet. Both fleet types use scaling policies to ensure there are enough streaming instances available when your users want to stream, while letting you (the admin) choose between application launch time, and cost. When your user launches their application on an Always-On streaming instance, the application starts loading almost immediately, while an On-Demand streaming instance has a brief wait (up to 90 seconds) while the streaming instance is powered on and made available. With Always-On streaming instances, you pay the hourly fee for the instance even when no user is streaming. With On-Demand streaming instances, you pay a lower stopped instance fee when the user isn’t streaming, and the hourly fee when they are.
The Amazon AppStream 2.0 Pricing Tool provides only an estimate of your AWS fees related to your usage of AppStream 2.0 and doesn’t include any taxes that might apply. Your actual fees depend on a variety of factors, including your actual usage of AWS services.
- Download the AppStream 2.0 Pricing Tool: Microsoft Excel File | OpenOffice Calc File
- Collect your actual or estimated concurrency usage by hour
How to use the AppStream 2.0 Pricing Tool
The AppStream 2.0 Pricing Tool is provided as a Microsoft Excel or OpenOffice Calc spreadsheet that enables you to enter in basic information about your usage, then provides a cost estimate for the AppStream 2.0 environment. The spreadsheet has a Price Estimator worksheet, and a Usage Pattern worksheet. Within the spreadsheet, the cells that require your input are denoted by a light blue background. The cells with a gray background are informational or aggregations, while the cells with green backgrounds are calculated cost estimates.
To get started, open the AppStream 2.0 Pricing Tool using Microsoft Office or Open Office. On the Price Estimator worksheet enter the following information:
- Cell B8: The total number of unique users that will stream in a given month. You will fill out expected concurrency per hour later.
- Cell B9: The Operating System to be used
- Cell B10: The AWS Region that will be used
- Cell B11: The instance type and size that will be used by your users
- Cell B12: The RDS SAL type – you can select between Commercial License Included, Bring Your Own License (BYOL) if you have RDS CALs that can be ported, or Academic License Included for qualified schools, universities, and public institutions. RDS SAL types are only applicable for Microsoft Windows operating system. For Amazon Linux 2, there is no RDS SAL type and you must set it to “Not Applicable”.
- Cell B13: Enter in a percent of concurrency that you want to maintain as buffer capacity to ensure new users are able to start streaming without waiting for new instances to be provisioned.
Note: Each cell has text that appears when you mouse over it providing additional details and instructions.
Once you have entered all of the information on the Price Estimator worksheet, switch to the Usage Pattern worksheet. On the Usage Pattern worksheet, enter the following:
- Cell C2: Enter the number of weeks per month. A default of 4 weeks per month has been entered.
- Cell C5: Enter in the number of days per work week. A default of 5 days per work week has been entered.
- Cell C7-C30: Fill in your actual or estimated concurrency per hour during a week day. If you expect no usage in specific hours, enter 0. Column D will be automatically calculated based on the buffer capacity entered on the Price Estimator worksheet.
- Cell C35: Enter in the number of days per weekend. A default of 2 days per weekend has been entered.
- Cell C37-C61: Fill in your actual or estimated concurrency per hour during a weekend day. If you expect no usage in specific hours, enter 0. Column D will be automatically calculated based on the buffer capacity entered on the Price Estimator worksheet.
The AppStream 2.0 Pricing Tool will automatically calculate the number of buffer instances based on what was entered in column C and the buffer capacity percent specified on the Price Estimator worksheet.
After you have filled in your per-hour usage patterns on the Usage Pattern worksheet, switch back to the Price Estimator worksheet. The AppStream 2.0 Pricing Tool calculates the effective monthly used and buffer hours, total monthly cost estimate, annualized cost estimate, and effective monthly cost per user estimate. The AppStream 2.0 Pricing Tool also calculates the cost savings by using On-Demand fleet type versus the Always-On fleet type, allowing you to choose between application launch speed and cost to deliver applications.
You can use the effectively monthly hours per user calculation to spot check whether the values entered in the usage pattern match your expectations of user usage. If they don’t, switch to the Usage Pattern worksheet, and adjust the per-hour usage pattern.
Now let’s walk through some examples of using the AppStream 2.0 Pricing Tool.
Example scenario streaming line of business applications
Suppose you are an enterprise that uses an application streaming technology to stream the SAP GUI to your users, or you’re a financial ISV that wants to stream your accounting software to your customers as a service and need to estimate the price of using AppStream 2.0. Your infrastructure is built in the Oregon (us-west-2) AWS Region, and you have 2,000 total users that will stream during the month. The application runs perfectly using the stream.standard.medium instance type and size. Your usage is fairly predictable during the weekday, so you maintain a 20% buffer, with sporadic usage over the weekend. Your enterprise has Microsoft License Mobility, enabling you to port your RDS CALs to AWS for use with AppStream 2.0. For users covered by your RDS CALs, you won’t incur monthly user fees. Your inputs on the Price Estimator worksheet will look like this:
You expect your usage pattern to be the following:
Based on this usage pattern, and the environment details, the AppStream 2.0 Pricing Tool estimates that the effective monthly cost per user would be $7.49 for Always-On, or $6.56 for On-Demand. You can find that at the bottom of the Price Estimator worksheet:
Example scenario streaming graphics apps
Now suppose you are an industrial engineering company that wants to stream GPU-intensive applications like SOLIDWORKS or Siemens NX to your design engineers who are located in France, Germany, and Italy. Currently you are deploying engineering workstations, but your engineers want portability in how and where they work by using ultrabooks. You want to estimate the cost of using AppStream 2.0 GPU-backed instances to augment providing your design engineers with ultrabooks. Your product lifecycle management and product data management infrastructure are in the Frankfurt (eu-central-1) AWS Region, and you have 3,000 design engineers who will stream sporadically throughout the day. You are going to use the stream.graphics-design.2xlarge for your design engineers to ensure they have sufficient CPU, memory, and GPU for them. You don’t have existing RDS CALs to port, so you will be buying them from AWS. You’re not too sure on your usage pattern, and so you want to maintain extra buffer capacity of 40% during the onboarding phase, then adjust after usage is identified. On weekends, you don’t expect any usage, but want to maintain some capacity for the design engineers who may need it. On the Price Estimator worksheet, you would enter:
And for the usage pattern, you expect:
Based on this usage pattern and the environment details, the AppStream 2.0 Pricing Tool estimates that the effective monthly cost per user per month would be $76.36 for an Always-On fleet, or $56.30 for an On-Demand fleet. In this scenario, you can save 26.3% with your users waiting up to 90 seconds for their application to launch with an On-Demand fleet. You can find this at the bottom of the Price Estimator worksheet:
Other costs to consider
The AppStream 2.0 Pricing Tool helps you with estimating your streaming instances costs. Other costs that are common to AppStream 2.0 environments are storage. For example, you will pay Amazon S3 storage fees if you enable AppStream 2.0’s Home Folder or Application Settings Persistence features for your users. Or, if you create an Amazon FSx file share for your users to save their data to.
AppStream 2.0 enables you to migrate away from your designed-for-peak-capacity application streaming environment to a dynamic, scalable, managed service where you only pay for what you use, when you use it. And with the AppStream 2.0 Pricing Tool, you can quickly estimate the cost per user to compare against your alternative options.