VMware Tanzu Application Service 6.0 Brings GenAI Updates and Boosts Developers and Platform Engineer Efficiencies

April 8, 2024

Written by Nick Kuhn and Mia Villarreal

We are excited to announce that VMware Tanzu Application Service 6.0 is now generally available! This latest release brings even more optimizations and integrations to the Tanzu Application Service platform that have been highly requested by our customers, such as GenAI capabilities, a smaller platform footprint option, improved security postures, and data protections. Tanzu Application Service Platform continues to provide enterprises and organizations with a turnkey, integrated modern application platform that deploys at the speed of business reliably across clouds with little to no downtime.

The release of Tanzu Application Service 6.0 encompasses the Tanzu ‘5 S’ framework that includes speed, stability, scalability, security, and savings. Customers leveraging Tanzu Application Service as their platform of choice continue to reap the benefits of cost savings with even more features that allow development and platform engineering teams to leverage GenAI, an option for a reduced footprint, observability, and easier operations.

Let’s take a look at the features included in Tanzu Application Service 6.0, now generally available.

Improved Developer Experience with GenAI and App Autoscaling Enhancements

Every release of Tanzu Application Service continues to improve the incredible developer experience of the platform. This release will focus on how Tanzu Application Service is rapidly evolving with Generative AI (GenAI) and improving the overall App Autoscaler experience.

Updates to GenAI for Tanzu Application Service (beta)

As GenAI takes the industry by storm, the Tanzu team has rapidly scaled a beta offering of GenAI for Tanzu Application Service. GenAI for Tanzu Application Service is a new tile, currently in beta, and aims to rapidly enable developers with GenAI functionality without having to learn how to deploy Private AI infrastructure or change existing development practices. 

Some new features released in the beta program include:

  • Improved vSphere Support: Platform engineers can now deploy large language models on larger GPUs, such as the Nvidia A100, on VMware vSphere-backed deployments. This allows for larger models to be deployed and for improved inferencing times

  • Azure Support: GenAI for Tanzu Application Service, has been validated to work with Azure as a support IaaS destination in addition to AWS, GCP, and vSphere, providing customers with more choices across cloud providers

  • Multi-Model Support: GenAI for Tanzu Application Service now supports multiple worker models running simultaneously, which expands the types of applications that can use this service. This enables platform engineers to deploy multiple types of LLMs optimized for different use cases. For instance, a platform engineer can deploy a model optimized for chatbots and models optimized for code completion

  • Improved inferencing speeds with vLLM support: In addition to the fastchat deployment option, platform engineers can serve LLMs via vLLM, optimized for inference throughput. This improves overall model response times and preparation for running models in production

Interested in testing it out for free? Sign up for the GenAI for Tanzu Application Service beta program to get access to these features!    

Improved Autoscaling based on Application CPU Usage

Application teams can save more time and effort in the long run by managing CPU-based autoscaling rules for their apps with less tuning and fewer errors.

App Autoscaler has been updated to use CPU entitlement metrics, which greatly improves the overall autoscaling of applications based on CPU usage. App Autoscaler provides a new rule that monitors an app’s average usage of its CPU entitlement and scales the app up or down accordingly. 

Unlike with the previous CPU processor utilization rule, the scaling thresholds for this rule do not need to change when developers adjust the memory allocation for the app or when the platform engineers adjust the size of the underlying Diego cell container-host VMs. When configured through the improved autoscaling UI in Apps Manager, developers get sensible default values of 30% and 80% for the low and high scaling thresholds, which should be a good starting point for most apps. Developers should have an easier and more intuitive experience configuring this new autoscaling rule for their apps. 

Improved Platform Footprint and Enhanced Operations 

Tanzu Application Service 6.0 is packed with operational efficiency improvements for platform engineers, enabling them to reduce the size of their infrastructure consumption and spend less time managing the platform. 

  • Small footprint now supports an in-place upgrade to a full Tanzu Application Service Installation: Small footprint for Tanzu Application Service now supports customers that want to upgrade to a “full” Tanzu Application Service installation. Small footprint Tanzu Application Service was generally considered to be used for proof of concept or sandbox environments, but this upgrade path will allow platform engineers to migrate to full installation without having to reinstall the platform. This should allow organizations to start with an overall smaller footprint (~4-6 virtual machines) at a lower cost and gives them the option to grow their platform footprint in their own time. The small footprint for Tanzu Application Service is also fully supported for production use and can be utilized in locations with limited infrastructure footprint, such as an edge location. 

  • Availability Zone Aware routing for Gorouter: The Gorouter has a new, opt-in feature to prefer local routing within an Availability Zone. When this option is enabled, the Gorouters will attempt to route all application traffic to application instances running on Diego cells within their local availability zone. This will improve the overall network traffic path and reduce overall round-trip latency to improve application performance. Gorouters will prefer not to send traffic to traffic to application instances running on other Availability Zones, but will in the event that no application instances are running in their zone.

  • Scale your dopplers and traffic controllers all the way to 0: For Tanzu Application Service installations using aggregate syslog drains to send app logs and metrics via the syslog protocol, platform engineers can now completely turn off the doppler and traffic controller VMs at the core of the loggregator subsystem. This optimization represents another step forward in the transition towards the new, more efficient “shared-nothing” observability architecture that is incrementally replacing earlier generations of observability subsystems, and helps platform engineers further reduce their infrastructure costs. Note: To completely scale down to 0, platform engineers must stop the use of firehose nozzles and other tiles that utilize nozzles, such as Healthwatch and App Metrics. 

  • Human-readable foundation names in the platform log stream: Platform engineers can configure a human-readable foundation tag in their outgoing platform log stream. This will improve the overall experience when sending multiple foundation log streams to a log aggregation platform, as platform engineers will be able to identify what foundation is sending logs, ultimately avoiding mistakes and eliminating the need to re-work.     

  • Certificate Rotation Dashboard for internal certificate management: Tanzu Operations Manager has been updated to include a Certificate Rotation Dashboard to continuously improve the internal platform certification rotation and management process. The dashboard allows platform engineers to track certification rotation status to easily understand the next step for the certificate rotation process within their Tanzu Application Service environments.

  • Configurable App File descriptor limits: Tanzu Application Service allows platform engineers to change the default Unix file descriptor limits for applications running on the platform. This feature enables applications to move to the platform that requires larger than normal numbers of Unix file descriptors. This allows for more application types to be migrated onto the Tanzu Application Service platform. 

Enhancements to Enable Improved Platform Security Postures

Security is top of mind for enterprises as the threat landscape changes daily. Tanzu Application Service is committed to enabling platform engineers with an improved platform security posture with these new features:

  • Early access to FIPS-compliant stemcells: Tanzu Application Service now supports the use of FIPS-compliant Jammy stemcells through early access to federal customers for their TAS foundations. Please reach out to the Tanzu team if interested in testing out this feature.

  • Manage local User Account and Authentication (UAA) Password policies when external identity providers are configured: Local UAA password policies can be set and maintained in conjunction with an external identity provider. This change enables better security posture for platform engineers who are managing local platform accounts.   

Tanzu Application Service Ecosystem Optimizations

Enhancements to the AWS Cloud Service Broker

Development teams and platform engineers running their Tanzu Application Service installations on top of AWS will benefit from the following features:

  • Support for the AWS Simple Queue Service (SQS): Platform engineers can now add production-ready plans for the AWS SQS service to their Tanzu Application Service marketplace offerings. Developers can easily deploy and use the AWS SQS service to expand upon their application use cases just like any other service supported by the Cloud Service Broker.

  • Support for AWS GovCloud Regions: The Cloud Service Broker for AWS is now supported for use in AWS GovCloud Regions. Customers who wish to use these regions can now begin to migrate off the legacy AWS service broker and take advantage of all the new features and functionality of the Cloud Service Broker, allowing them with more options across cloud providers.

Download Tanzu Application Service 6.0 on Tanzu Network today! Join us at Cloud Foundry Day North America on May 15th, 2024.

VMware makes no guarantee that services announced in preview or beta will become available at a future date. The information in this press release is for informational purposes only and may not be incorporated into any contract. This article may contain hyperlinks to non-VMware websites that are created and maintained by third parties who are solely responsible for the content on such websites.

Previous
What Platforms (Should) Do for Developers - things to ask the Platform Engineering group
What Platforms (Should) Do for Developers - things to ask the Platform Engineering group

If you're lucky enough to have platform group in your company, @thecote shares three things you should ask ...

Next
7 Ways to Improve the Developer Experience and Accelerate Software Delivery on Kubernetes
7 Ways to Improve the Developer Experience and Accelerate Software Delivery on Kubernetes