Adapt the size of your AppStream fleet dynamically

AppStream Optimiser

What can we do for you

AppStream Optimiser

AppStream 2.0 logo

Amazon AppStream 2.0 is a fully managed non-persistent application and desktop streaming service. You manage your desktop applications centrally on AppStream 2.0 and deliver them to any computer securely. You can easily scale to any number of users across the globe without acquiring, provisioning, and operating hardware or infrastructure. 

AppStream charges largely depend upon the fleet size of instances. Such a fleet can be offered under two operational modes: always on and on demand. For always on fleets you pay for the provisioned fleet whether it is actually used or not. For on demand fleets instances you are charged at different rates whether they are being used or not. The used instances are charged their normal hourly fee rate and the unused ones are paused and are charged at a lower hourly fee rate.

Whether you use on demand or always on instances, the better you match AppStream demand to the fleet size the lower your usage and the lower your bill becomes. In either case managing the fleet size with the standard scaling options is challenging.

Logo of certified AWS Advanced Consulting Partner for Cambrian Technologies.
Why choose us?

A trusted partner

Scaling the AppStream fleet with default AWS tooling is based on current utilization or a schedule. In both cases you have to set parameters based on estimations and assumptions of the usage patterns. And however you manage this, there are significant downsides:

  • you may select to overprovision capacity to ensure AppStream availability, resulting in large excess costs for unused capacity;
  • if you do not provision with a large safety margin some users will face significant delays for AppStream to become available when demand increases rapidly (start of day; change of shift etc.);
  • you cannot accommodate unforeseen sudden peaks in demand and this has to be adjusted when noticed (while users are waiting);
  • managing unforeseen changes in demand results in the cumbersome task of adjusting scaling parameters, and later re-adjusting these to default values;
  • you pay for unused capacity if users stop using an AppStream session, but do not terminate the session properly (closing the laptop screen or just closing the browser tab where AppStream is running).

Either way this results in paying more than you actually use and requires manual management of the fleet. This problem is especially relevant with a larger number of users because their individual usage patterns do not even out in the bill.

The AppStream Optimiser provides a fully automated solution for the management of the AppStream 2.0 fleet achieving a utilization that is impossible to realize with the standard AWS tooling. In addition, you no longer need to put effort in managing the appstream fleet.

Cambrian Technologies enables you to adapt the size of your AppStream fleet dynamically to match demand by analyzing the usage history and simultaneously inspecting today’s trends in real time. This will increase utilization typically above 90% and is suitable for critical fleets where insufficient capacity errors should be zero. It will also keep this level of utilization while usage patterns change, either gracefully (holidays) or rapidly (as a result of unexpected events such as Covid-19, where people change their working hour patterns). Due to a constant changing of parameters capacity can be increased and decreased rapidly or gradually depending on what demand requires.

Below is a graphical representation of how demand usage looks before the AppStream Optimiser has been implemented.

How it works

The Cambrian Technologies AppStream Optimiser is a fully managed service that automatically scales your AppStream fleet to a utilization of >90%. It uses machine learning to predict usage patterns and quickly scales the fleet no matter the size. This avoids end users having to wait for a new appstream to be launched while simultaneously realizing typical cost savings of 20% to 40%.

Additionally the model can be trimmed to specific customer needs, adjusting margins of safety, adjusting scale down settings upon inactivity based on the user profile(s) and accommodating for country and company specific holidays.

Additional product features:

  • Intelligently anticipates Amazon AppStream pricing model characteristics. For instance scaling down is not performed during the first hour as this is charged regardless of usage.
  • Scaling is performed based on real time usage as well as a historical usage. Machine learning modeling ensures auto scaling is adapted and optimized continuously.
  • Scaling incorporates foreseen (weekdays, holidays) and unforeseen usage patterns (sudden capacity peak) because scaling is performed rapidly and based on real time actual usage.
  • Scaling is performed automatically, manual AppStream fleet management is no longer required, the model performs scaling for you.
  • Detects and acts upon user inactivity. If a user stops using AppStream without terminating the session, the AppStream session continues running. The advanced settings can be used to automatically shut down inactive instances based on customers preference (time of day, user etc.).
  • The service includes monitoring tooling to verify effectiveness of the model monitoring actual usage and utilization.
  • Suitable and relevant for all pricing models: on-demand, always-on or a mix.
  • Standard reporting enables insight into application and appstream usage. This can help a company optimise the software licenses costs by monitoring actual application usage.
  • Custom reporting is available to track concurrent and individual usage specific to applications used via AppStream. Includes reporting to help management to ensure correct management of sessions by users, which will reduce idle sessions.
Key activities
  1. Intake & Access
    Identify customer needs and preferences (safety margins, company holidays and scaling down preferences are identified) and grant required access to AppStream data.
  2. Configure & train
    Historical usage data of the appstream fleet is released to the solution, the model trained and customer specific parameters implemented.
  3. Testing
    Testing is performed to validate outcomes, before implementation.
  4. Implementation
    The AppStream Optimizer is implemented enabling auto scheduling.
  5. Monitoring and continuous improvement
    Once the automation is implemented the effectiveness is monitored continuously and you benefit from ongoing model optimizations.
  6. Reporting
    Next to standard reporting which comes with the tool, custom reporting can be designed and implemented.
Customer contribution

Stakeholder involvement
Involvement in the initial call, discovery workshop to determine preferences, and validation of the outcome discussions.

Grant (AWS cross-account) access to the AppStream usage history (an S3 bucket) and AppStream management of the fleets optimized.

Reporting needs
Input on reporting requirements to adjust reports (if relevant).

Benefits of AppStream Optimiser

Realize significant cost saving by increasing fleet utilization (>90%)

No more waiting time for end users for instances to launch

Scalability is performed no matter the size, appstream fleet management is fully automated

Gain insight in AppStream usage & utilization based on standard reporting

Gain insight in application usage & utilization to optimize license costs, monitor compliance based on standard and custom reports

Scroll to Top