Unexplored Territory

Brandon Frost shares his experience with memory tiering and the use of Intel Optane in his large private cloud environment. He explains how memory tiering helped optimize memory usage and reduce costs by offloading cold memory to a non-active tier. He also discusses the challenges of upgrading memory and the benefits of using NVMe-based tiering. Brandon expresses his interest in Project Peaberry, a memory tiering offloading card, and its potential use cases. He highlights the importance of his team in developing capacity data gathering and analytics solutions.

Takeaways:

• Memory tiering helps optimize memory usage and reduce costs by offloading cold memory to a non-active tier.
• Using NVMe-based tiering provides flexibility and scalability in memory allocation.
• Project Peaberry, a memory tiering offloading card, offers potential benefits in terms of performance and cost-effectiveness.
• Effective capacity planning requires robust data gathering and analytics solutions.
• Brandon Frost emphasizes the importance of his team in developing and implementing memory optimization strategies.

Links:

• VMware Explore session by Brandon and Arvind on memory tiering
• Memory Tiering blog on vmware.com

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

VMware Explore 2024 in Las Vegas is a wrap. The team was present and will give you the opportunity to get an idea what the show was like. You will hear large number of guests covering some of the announcements, some specific aspects of the conference (like the Hands-on Labs), but also some of the vendors on the expo. The episode contains the following topics:

• The future of VMUG with Brad Tompkins
• A hands-on labs tour with Brandon Bazan and Joshua Schnee
• Spring One and Spring AI with DaShaun Carter
• VCF and AI at VMware Explore with Stephen Wagner
• EUC and 10zig with Kevin Greenway
• Google Cloud with Simon Long
• Microsoft AVS with Rene van den Bedem
• OVH Cloud with Pascal Jailon
• Pure & Stretched vVols with Cody Hosterman

If you like to watch one of the sessions mentioned, check out William Lam's Github page:

https://github.com/lamw/vmware-explore-2024-session-urls/blob/master/vmware-explore-us.md

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

Arvind Jagannath, a platform product manager at Broadcom, discusses memory tiering and its benefits for customers with Duncan. He explains that enterprise applications are becoming more memory-bound and require better performance. Memory tiering helps address the CPU to memory imbalance that many customers face, allowing for better CPU utilization and reducing costs. Arvind also mentions the challenges of expanding memory and the need for a solution like memory tiering. He introduces Project Capitola, which has evolved into NVMe-based memory tiering, and discusses the benefits and recommendations for using NVMe devices. Arvind also touches on the integration of memory tiering with advanced vSphere features like DRS and HA, and hints at future developments with Project Peaberry and CXL. 

Takeaways

• Enterprise applications are becoming more memory-bound and require better performance.
• Memory tiering helps address the CPU to memory imbalance and improves CPU utilization.
• NVMe-based memory tiering provides additional memory capacity and reduces costs.
• Integration with advanced vSphere features like DRS and HA is being developed.
• Future developments include Project Peaberry and the use of CXL for intelligent devices.

Links

• Explore Session - Memory Tiering: Power Your Server Infrastructure With Memory Innovations [VCFB1198LV]
• Yellow-Bricks.com - Memory Tiering Blog
• SNIA Recording on Memory Tiering and CXL accelerators 

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

In this conversation, Katarina Brookfield discusses her career trajectory and her current role at Broadcom. She shares defining moments in her career, including her experience working on the Black Sea Maritime Archaeology project. The conversation then shifts to the newly announced vSphere IaaS control plane and its benefits. Katarina explains that the control plane provides a comprehensive solution for deploying workloads, including additional services like storage provisioning and load balancing. The conversation also covers the self-service nature of the control plane, the different interfaces for consumers and admins, and the integration of HashiCorp Packer for building and customizing VM images. The TKG service, which allows for the deployment of managed Kubernetes clusters, is also discussed, highlighting its ease of use and integration with vSphere. The conversation concludes with a discussion of the new features in the latest version of the TKG service, including cluster auto-scaling and the decoupling of TKG from vCenter.

Takeaways

• The vSphere IaaS control plane provides a comprehensive solution for deploying workloads, including additional services like storage provisioning and load balancing.
• The control plane offers a self-service experience for consumers, allowing them to easily deploy the services they need.
• Different interfaces, including APIs, CLI, and UI, cater to the preferences of different users, making it accessible to both admins and consumers.
• The integration of HashiCorp Packer allows for the building and customization of VM images, providing flexibility and automation.
• The TKG service simplifies the deployment of managed Kubernetes clusters, making it accessible to users with little Kubernetes experience.
• The latest version of the TKG service decouples it from vCenter, allowing for faster delivery of new Kubernetes versions.
• New features in the TKG service include cluster auto-scaling and the integration of HashiCorp Packer for building and customizing VM images.

Chapters

00:00 - Kat's Career Trajectory and the Role of Defining Moments

09:20 - The Comprehensive Solution of the vSphere IaaS Control Plane

11:02 - Enabling Self-Service and Catering to Different User Preferences

18:14 - Flexibility and Automation with HashiCorp Packer Integration

22:47 - Simplifying Kubernetes Deployment with the TKG Service

29:14 - Decoupling TKG from vCenter for Faster Delivery of Kubernetes Versions

38:36 - New Features in the Latest Version of the TKG Service

Links

Check Katarina's sessions at VMware Explore in Las Vegas here

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

Summary:

Ala Dewberry introduces Project Marigold, a plugin that aims to make GPUs a first-class citizen on vSphere by providing visibility and observability of GPU metrics at the cluster and data center level. The plugin helps infrastructure administrators manage GPUs more strategically and enables data scientists and application developers to have a frictionless experience. Ala also discusses the use cases for private AI and encourages users to give feedback on the plugin.

Takeaways

• Defining moments in our careers can shape our trajectory and open up new opportunities for growth and learning.
• Project Marigold aims to make GPUs a first-class citizen on vSphere by providing visibility and observability of GPU metrics at the cluster and data center level.
• The plugin helps infrastructure administrators manage GPUs more strategically and enables data scientists and application developers to have a frictionless experience.
• Private AI allows organizations to leverage their data centers for AI workloads that require data privacy, security, and compliance.

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

VMware recently released vSphere and vSAN 8.0 Update 3, and of course, we had to invite Feidhlim, Jason, and Pete back on the show to discuss what's new in these releases. There's awesome new functionality released and some great enhancements, so make sure to listen to the full episode.

Key Takeaways:

• vSphere 8.0 update 3 introduces the vSphere Live Patch Update Path, which allows for patching ESXi hosts without evacuating VMs or entering full maintenance mode.
• Improvements in GPU functionality include the ability to use two DPUs in an ESXi host for availability, better support for VGPUs with different profiles and memory sizes, and simplified activation of GPU mobility with DRS.
• The vSphere cluster service (VCLS) has been re-architected to reduce resource consumption and improve rollback mechanisms.
• The 8.0 update 3 introduces stretched VVols, which customers have been asking for, and support for stretched fault tolerance.
• There are enhancements in VVols, including unmapped support for NVMe over fabrics.
• The updates in NVMe over Fabrics provide faster data migration and cloning.
• NFS enhancements include VMK port binding and support for NFS version 4.1.
• vSAN 8.0 U3 introduces new features and enhancements in flexible topologies, agile data protection, and enhanced management.
• The support for stretch cluster arrangement in VCF allows customers to take full advantage of ESA and improve performance, storage efficiency, and resilience.
• The full support of vSAN Max as principal storage within a workload domain enables customers to maintain a centralized shared storage model while leveraging the capabilities of vSAN.
• vSAN data protection allows users to create snapshots based on groups of VMs, set snapshotting schedules, and easily recover VMs without them being part of the inventory.
• Enhancements in alerting capabilities for NVMe storage devices and proactive hardware management provide better visibility and intelligence about the health and wellbeing of storage devices.

Follow us on Twitter for updates and news about upcoming episodes: https://twitter.com/UnexploredPod.

Last but not least, make sure to hit that subscribe button, rate wherever possible, and share the episode with your friends and colleagues!

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

Summary

The conversation explores the features and benefits of the Tanzu Platform, which accelerates application delivery and simplifies the development process. The platform abstracts the complexities of Kubernetes, allowing developers to focus on writing code and creating business value. It offers four golden commands: build, bind, deploy, and scale, which streamline the development process. The platform also reduces administrative overhead for IT teams, with a developer-to-platform engineer ratio of up to 1,000 to 1. The Tanzu Platform is used in various industries, including automotive, finance, healthcare, and retail, to develop and run mission-critical applications.

Keywords

Tanzu Platform, application delivery, Kubernetes, developer experience, administrative overhead, use cases

Takeaways

• The Tanzu Platform accelerates application delivery and simplifies the development process.
• It abstracts the complexities of Kubernetes, allowing developers to focus on writing code.
• The platform reduces administrative overhead for IT teams, with a high developer-to-platform engineer ratio.
• The Tanzu Platform is used in various industries to develop and run mission-critical applications.

Sound Bites

• "The safest developer experience must also be the easiest."
• "You can repair very quickly and even in an automated way without any manual intervention or without even downtime."
• "Imagine doing this without golden paths. It's not practical to try and repave if you have hundreds of different ways of deploying your apps to production."

Chapters

00:00
Introduction and Background

07:20
Overview of the Tanzu Platform

12:21
The Develop, Operate, Optimize Philosophy

28:00
Integration with the VCF Platform

37:27
Use Cases of the Tanzu Platform

50:04
Impact on AI Workloads

53:25
Closing Remarks

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

In this conversation, Duncan and Cormac Hogan discuss VMware's Data Services Manager (DSM) and its role in offering data services in a full-stack private cloud. They cover topics such as the use cases for DSM, the integration with Kubernetes, the support for different databases, the automation capabilities, and the licensing model. Cormac highlights the features of DSM, including lifecycle management, backups, scaling, monitoring, and advanced settings. He also mentions the upcoming release of new features and additional data services. 

Takeaways

• Data Services Manager (DSM) is a VMware product that offers data services in a full-stack private cloud.
• DSM integrates with Kubernetes and allows VI administrators to maintain control of vSphere resources while offering data services.
• DSM supports databases such as Postgres and MySQL, with support for other data services like AlloyDB in tech preview.
• DSM provides features such as lifecycle management, backups, scaling, monitoring, and advanced settings.
• DSM is included in VMware Cloud Foundation (VCF) and support can be added through the Private AI Foundation add-on.

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

In this conversation, Myles Gray discusses the AI workflow and its personas, the responsibilities of data scientists and developers in deploying AI models, the role of infrastructure administrators, and the challenges of deploying models at the edge. He also explains the concept of quantization and the importance of accuracy in models. Additionally, he talks about the pipeline for deploying models and the difference between unit testing and integration testing. Unit testing is used to test the functionality of a single module or function within an application. Integration testing involves testing the interaction between different components or applications. MLflow and other tools are used to store and manage ML models. Smaller models are emerging as a solution to the resource constraints of large models. Collaboration between different personas is important for ensuring security and governance in AI projects. Data governance policies are crucial for maintaining data quality and consistency.

Takeaways

• The AI workflow involves multiple personas, including data scientists, developers, and infrastructure administrators.
• Data scientists play a crucial role in developing AI models, while developers are responsible for deploying the models into production.
• Infrastructure administrators need to consider the virtualization layer and ensure efficient and easy consumption of infrastructure components.
• Deploying AI models at the edge requires quantization to reduce model size and considerations for form factor, scale, and connectivity.
• The pipeline for deploying models involves steps such as unit testing, scanning for vulnerabilities, building container images, and pushing to a registry.
• Unit testing focuses on testing individual components, while integration testing ensures the compatibility and functionality of the entire system. Unit testing is used to test the functionality of a single module or function within an application.
• Integration testing involves testing the interaction between different components or applications.
• MLflow and other tools are used to store and manage ML models.
• Smaller models are emerging as a solution to the resource constraints of large models.
• Collaboration between different personas is important for ensuring security and governance in AI projects.
• Data governance policies are crucial for maintaining data quality and consistency.

Chapters

• 00:00 Understanding the AI Workflow and Personas
• 03:24 The Role of Data Scientists and Developers in Deploying AI Models
• 08:47 The Responsibilities of Infrastructure Administrators
• 15:25 Challenges of Deploying Models at the Edge
• 20:29 The Pipeline for Deploying AI Models
• 24:45 Unit Testing vs. Integration Testing
• 28:22 Managing ML Models with MLflow and Other Tools
• 32:17 The Emergence of Smaller Models
• 39:58 Collaboration for Security and Governance in AI Projects
• 46:32 The Importance of Data Governance

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.

Summary
For this special edition of the podcast Duncan invited Michael Roy to discuss the latest VMware Workstation and VMware Fusion announcements. VMware Workstation and Fusion are desktop hypervisor products that allow users to run virtual machines on their PC or Mac. Starting today, Workstation and Fusion commercial licenses will only be available through annual subscriptions. The price for both products is now $199 per year. The free versions of Fusion Player and Workstation Player are being discontinued, but the Pro versions will be available for free for personal use. Support for personal use products will be community-based, while commercial users will have support included in their subscription. The focus of future innovation will be on the integration between vSphere and Workstation/Fusion, providing a local virtual sandbox for learning, development, and testing.

Takeaways

• VMware Workstation and Fusion are desktop hypervisor products for running virtual machines on PC and Mac.
• Commercial use of Workstation and Fusion is shifting from perpetual licenses to annual subscriptions.
• The free versions of Fusion Player and Workstation Player are being discontinued, but the Pro versions will be available for free for personal use.
• Support for personal use products will be community-based, while commercial users will have support included in their subscription.
• Future innovation will focus on integrating vSphere with Workstation and Fusion to provide a local virtual sandbox for learning, development, and testing.

Links

• Announcement Blog
• The Register article

Disclaimer: The thoughts and opinions shared in this podcast are our own/guest(s), and not necessarily those of Broadcom or VMware by Broadcom.