Azure Stream Analytics has Launched a New Competitive Pricing Model!
Published May 23 2023 08:00 AM 5,374 Views

Our customers already enjoy ASA’s exceptional performance, high availability, low-code/no-code experiences, and rich development tools.  To make streaming analytics with ASA an even easier win, we are announcing a new competitive pricing model which offers price reductions of up to 80% with no changes to ASA’s full suite of capabilities. 

 

This model supports the team’s mission of democratizing stream processing for all. 

 

Read on to learn about:

  • What the overall price reduction is and how discounted tiering works.
  • How ASA measures compute, represented as Streaming Units or SUs.

 

Best in class price to performance ratio:

To set new pricing, we dug deep to set the most competitive price we could.  In doing so, we are proud to announce a best-in-class price to performance ratio in the streaming world, with no changes to ASA's full suite of built-in capabilities.

 

Before we dive into pricing details, it’s important to understand that ASA’s SU structure with V2 pricing is different.  If you’re not familiar with Streaming Units, or SUs, they represent the computing resources which are allocated to execute a Stream Analytics job.  The higher the number of SUs, the more CPU and memory resources are allocated to your job.

The intention behind this change was to simplify the model both in terms of understanding usage and for scaling ASA jobs. 

 

In today’s model, every 6 SUs correspond to a single streaming node for a job.  Jobs may run with 1 and 3 SUs as well and these correspond with fractional streaming nodes.  As such, in ASA’s original model scaling occurs in increments of 6 beyond 6 SU jobs, to 12, 18, 24 and beyond by adding more streaming nodes that provide distributed computing resources.

 

In the V2 model, 1 SU V2 corresponds to a single streaming node for a job.  2 SU V2s corresponds to 2… 3 to 3… and so on.  The model has now become much simpler to understand (and explain), and the compute allocation remains the same.

 

Fractional compute resources continue to be supported for light jobs and POCs, represented at 1/3 and 2/3s.

 

The underlying compute from SU to SU V2 is as follows:

Anasheh_Boisvert_2-1689962789363.png

 

How discounted tiering works:

With the simplified SU structure, we offer graduated tiers for eligible customers based on a monthly aggregation model.  What this means is that as your usage of streaming units grow, the price per unit drops. 

 

Let’s look at pricing in East US as an example:

 

Anasheh_Boisvert_0-1689961037059.png

 

Starting price here represents a fractional streaming unit at 1/3, this is the lightest compute job you can have in ASA.  Light compute jobs are great for small workloads and POCs.

 

The price per unit within the tiers represents a full streaming node, and tiering is determined based on an accumulated monthly aggregation of streaming unit hours.

 

Here's an example of two jobs in West US with the same compute allocation.  One with ASA’s original pricing and one with V2 pricing.

 

 Original standard pricing:

Anasheh_Boisvert_1-1689961037065.jpeg

 

 

Standard V2 pricing:

Anasheh_Boisvert_2-1689961037067.jpeg

 

Note that these jobs have the same compute allocation thus same performance.  ASA’s realigned compute allocation aligns each streaming unit with 1 streaming node.  In the previous model, every 6 streaming units corresponded with a single streaming node.  

 

IMPORTANT

Please note that Azure pricing differs depending on what region you are in.  For pricing details in your region, please visit the Azure Stream Analytics pricing page

 

To compare original and V2 pricing, visit Azure Calculator and choose Azure Stream Analytics from the list of services.

 

Availability    

V2 pricing is available in all ASA regions.

 

Dedicated clusters

Dedicated cluster pricing follows the same discounting structure with a few changes.

Today, a minimum of 36 SUs is required for a cluster, this brings the minimum monthly cost for an ASA cluster to $2,891.80 (East US).  With the new pricing model, only a minimum of 12 SU V2s is required to stand up a dedicated cluster.  With the new pricing, a V2 dedicated cluster costs are now approximately $946 per month (East US).

 

Customers can create clusters and jobs same as today as there are no functionality changes.  Each ASA Cluster is provisioned with the same allocation as before, and from a capacity standpoint, cluster sizes do not change. 

 

Jobs in a dedicated cluster can run with SU V2s set to 1/3, 2/3, 1, 2, 3… and so on.  Existing dedicated clusters can be converted to V2. 

 

Please note that jobs in a dedicated cluster created with SU V2 capacity can only support jobs with SU V2.  Meaning, you cannot run both current and NEW SUs in a dedicated cluster.  Mix and match is not supported due to capacity complications.

 

IMPORTANT

Virtual Network Integration support is also available in Public Preview.  Virtual network (VNET) support enables you to lock down access to Azure Stream Analytics to your virtual network infrastructure.

Learn more here.

 

Onboarding steps

Azure Portal:

  1. Look for banner on ASA’s Overview page which reads ‘This job is eligible for Stream Analytics’ new pricing model.
  2. Click on link ‘To manage pricing, click here’ and follow the prompts to upgrade to ASA’s new pricing model.

Anasheh_Boisvert_3-1689961287745.png

 

Visual Studio Code:

  1. Open the JobConfig.json
  2. The old ‘StreamingUnits’ field is now auto replaced by a ‘Sku’ field with ‘Name’ and ‘StreamingUnits’ under it.
  3. Choose “StandardV2” for ‘Sku’ name. 
  4. Updating the ‘Name’ to “StandardV2” will trigger an auto update for ‘StreamingUnits’.   

Anasheh_Boisvert_4-1689961312664.png

 

Continuous Integration and Continuous Deployment (CI/CD):

  1. After `Compile Script` with `ARM Template V2` or `Bicep Template`, you’ll see the sku name and streaming units be correctly set in the template file.  This template file can be used in CI/CD.

Anasheh_Boisvert_5-1689961478043.png

 

Old SKU deprecation path

Azure Stream Analytics will continue to support both structures and pricing models for the foreseeable future so there are no business interruptions.

 

Quotas and limits

All quotes and limits will remain with a conversion from SU to V2.  For example, the existing SU limit per sub per region will change from 500 SUs to 84 SU-V2s.

 

Troubleshooting

If you encounter problems while migrating to the new structure, please follow the prompts and error handling scenarios which have been created as part of the onboarding process to help you self-service.

 

If you are unable to resolve the issues encountered, contact Microsoft Support.

 

Feedback

For direct feedback on ASA’s new pricing model, please contact us at askasa@microsoft.com.

 

3 Comments
Co-Authors
Version history
Last update:
‎Jul 21 2023 11:07 AM
Updated by: