posted

2 Comments

We know you have a lot of questions about Project Pacific, so to help you gain a better understanding, we hosted a vSphere Tweet Chat with some of our experts. Our two main featured guests were Tasha Drew and Jared Rosoff, but a variety of other experts, such as Joe Beda, Nikitha Suryadevara, David Stamen, Dilpreet Bindra, and Timothy St. Clair joined us as well. View their invaluable insights into the inner workings of Project Pacific, along with what’s in store by checking out the full vSphere Tweet Chat recap:

Q1: What is Project Pacific?

Tasha Drew

Jared Rosoff

A1: Project Pacific is a re-architecture of vSphere that puts Kubernetes at the heart of the platform. It lets you use Kubernetes to manage not just containers, but all of your virtual infrastructure. As a part of VMware’s Tanzu portfolio of products, Project Pacific provides the layer that RUNs modern applications. Tanzu also brings all of the tools you need to BUILD and MANAGE modern apps too. Altogether, it’s a powerful portfolio of solutions. When I explain Project Pacific to folks, I usually say that it is really 3 things:

  • First, it is embedding a special purpose k8s cluster in vSphere in order to make vSphere better. It exposes a direct interface using k8s API patterns.
  • Second, Project Pacific is a set of capabilities to make Kubernetes run great on top of vSphere. This includes exciting things like vSphere Native Pods, where we adapt ESXi to run Pods natively.
  • Third, Project Pacific is an extended set of capabilities (using the supervisor cluster) for launching “Guest Clusters” for running applications.

Q2: Why Kubernetes? And why does rearchitecting vSphere make sense?

A2: Kubernetes is awesome because it’s a widely adopted, fully open source API driven control plane that you can easily extend. It’s easy to build on top of, and since we’re building up from the infrastructure, that’s exactly what we, our users, developers, and partners need. As far as re-architecting vSphere, Kubernetes gives us the developer-focused interface we need to enable self-service and still gives Ops total control via the vSphere UI & tools they know and love.

Joe Beda

Nikitha Suryadevara

A2: Rearchitecting vSphere gives you the best of both worlds!

David Stamen

A2: Why Not? vSphere is already one of the best places to run your workloads, why not integrate one of the well-known tools your developers are already using.


Q3: How did the Heptio acquisition influence Project Pacific?

Tasha Drew

Joe Beda

A3: I have a bit of a unique perspective here. I think one of the things that Heptio did is bring the perspective around using Kubernetes as a distributed control plane vs. just focusing on running containers. This led to the supervisor vs. guest cluster split.

Dilpreet Bindra

A3: While we had been working on aspects of Project Pacific prior to the acquisition, we were only able to realize the full value of Pacific and Tanzu. Thanks to @forjared, @kostadis_tech and @jbed – they all worked to nail down the current strategy.

David Stamen

A3: Allowed two amazing teams to come together and work on a great product!

 


Q4: What about the future of VMs?

Tasha Drew

A4: Being able to automate the lifecycle of applications on VMs as an infra choice is 100% the future, and that’s what we can do with Project Pacific and Kubernetes! Now your devs get VMs on demand (via a VM Operator), your Ops team still has complete mgmt/control via vCenter policy, and you can use the Operator/CRD pattern to automate your application’s lifecycle on VMs and containers.

Joe Beda

A4: VMs and Containers are complimentary. The lines are blurring with things like vSphere Native Pods. And also having automated infra available makes managing clusters as cattle possible with things like Cluster API. Horses for courses.

David Stamen

A4: VMs will continue to exist, but know they can run alongside their best friends, the container!

Nikitha Suryadevara


Q5: Why is Project Pacific important for VI Admins?

Tasha Drew

A5: So many reasons!! You can create Namespaces in the vSphere UI, attach policy (compute, network, storage, security, availability…), and give devs access to it. Managing at a namespace/app level makes you way more powerful and unlocks cloud-like self-service to your devs.

Jared Rosoff

A5: VI admins are the lifeblood of enterprise infrastructure. They provide the virtualized compute, network and storage resources that run the majority of enterprise workloads today. With Project Pacific, VI admins extend their skills to manage ALL workloads, not just the VMs.The biggest challenge facing enterprise IT today is how to evolve their infrastructure architecture to support these new applications. Almost every solution out there before Project Pacific required that you build a new infrastructure platform and then refactor all your existing apps. With Project Pacific, we finally have an evolutionary path where your existing platforms, tools and skills can support these new applications. With little more than an upgrade of your existing infrastructure platform, you can turn your data center into a modern application platform.

Nikitha Suryadevara

A5: Kubernetes is so natively integrated into familiar screens and workflows in vSphere that the learning curve (if any) is going to be minimal.

Joe Beda

A5: Project Pacific gives VI admins new tools for application teams so they can be more effective. Also, there are more direct ways for those app teams to interact with vSphere.


Q6: Can you share some insight into how Kubernetes Namespace will influence app-focused management?

Tasha Drew

Jared Rosoff

A6: Modern applications are not just a single VM or container, they’re groups of VMs, Containers, Disks, and Networks that implement a service or application. We want to move infrastructure management to the point where admins can manage logical applications, not just VMs. Project Pacific embraces Namespaces as the new unit of infrastructure management. Tasks like resource allocation, mobility, security and data protection can work at the application level rather than having to be configured on each VM. This creates massive time savings for the admin. Remember that this is a layer UNDERNEATH regular Kubernetes clusters. So, you can create a namespace in vSphere, and then provision a bunch of k8s clusters into it and treat it as a single unit of management. It’s turtles all the way down.

Dilpreet Bindra

A6: Kubernetes Namespace becomes the fundamental management unit for your complex workloads in Project Pacific.

David Stamen

A6: Namespaces are awesome! They let you allocate specific resources such as CPU, Memory and Storage to make sure your applications can run smoothly!


Q7: What are the targeted Project Pacific use cases?

Tasha Drew

A7: Bring us your containerized cloud-native applications! Your legacy VM-based workloads! Your distributed stateful data services! Your FaaS!

Jared Rosoff

A7: We see a broad interest in Project Pacific for lots of different kinds of apps. Web applications, AI/ML, network function virtualization, internet of things, and stream processing are some common ones. There are two big trends we see across these apps. 1: Organizations that want to modernize their application development processes to be more agile and productive. 2: Organizations that have mission critical workloads that need the highest degrees of security and availability. One thing that I think is lost in this “pets vs. cattle” discussion is that if you’re a farmer, you might not be attached to an individual cow, but you DEFINITELY care whether or not your farm is producing milk. If you depend on your apps, vSphere is the best place to run them.

David Stamen

A7: Any application that can run as a container is welcome to run on Project Pacific!


Q8: What do I need to do to get ready for Project Pacific?

Tasha Drew

A8: We are going to be shipping training docs and making things super easy to get started on the enablement front from vSphere!

Jared Rosoff

A8: If you’re a VI admin today, you’re already 90% of the way there. If you’ve just been doing compute virtualization, now is the time to start learning about network and storage virtualization as these technologies will become ubiquitous with modern applications. You should be engaging with your development teams and working with them to develop new workflows and processes for application lifecycle management. The more involved they are in this process, the better your evolution will go. Of course, you should also sign up for the Project Pacific beta, here.

David Stamen

A8: If you are not familiar with @kubernetesio, its time to start learning! I recommend the Kubernetes Academy (https://kubernetes.academy) and @nigelpoulton‘s Pluralsight Series. Also, don’t’ forget to sign up for the beta!


Q9: What is your favorite capability that Project Pacific will offer?

Jared Rosoff

A9: Oh, it’s hard to pick a favorite child. The CRX is pretty exciting. The idea that you can get the same kinds of performance and simplicity of containers but with the security and performance properties of a virtual machine is pretty killer. I also love the fact that Kubernetes becomes the IaaS API. Having been an AWS user for more than 10 years, k8s based IaaS feels like what AWS should have been. Using declarative configuration that’s source control friendly is dope. Having it native to the platform is super dope.

Tasha Drew

A9: @Forjared Imma let you finish, but the vSphere Kubernetes Service + clusterAPI service + VM Operator is THE BEST FEATURE OF ALL TIME.

Nikitha Suryadevara

A9: Being able to manage all the different objects in a namespace (VMs, containers, native pods, Kubernetes clusters) transparently is an amazing capability offered by Project Pacific.

David Stamen

A9: I think my favorite capability is the ability to run native pods on vSphere. Think of all that performance!


Q10: Do I need Kubernetes to run Kubernetes? And does it involve turtles all the way down?

Tasha Drew

A10: You need all them turtles!! As you may know, one of the jokes in k8s is the best way to handle lifecycle of a k8s cluster is from … another k8s cluster. So, where do you get your first cluster?!

Jared Rosoff

A10: Do you NEED it? no. But once you try it, you’ll never go back. Running Kubernetes needs a programmatically controlled infrastructure that is able to orchestrate all of the components of a Kubernetes cluster – not just provisioning, but also health checks, recovery, upgrades… Sound familiar? Kubernetes was designed to orchestrate complex applications. So, turning it on Kubernetes itself is like chocolate and peanut butter. It just FEELS right. And technically it’s not turtles all the way down. There’s a ring 0 turtle that is the most privileged turtle.

David Stamen

A10: You do not need Kubernetes to run Kubernetes, that is what Project Pacific is for! The best turtles are the @VMwareTurtles.


Q11: How will Project Pacific impact VMware’s partner ecosystem?

Tasha Drew

A11: It provides a powerful, consistent way for partners to provide their services on top of or as part of vSphere, using the Project Pacific k8s control plane, exposing things as k8s objects, and/or running Operators/CRDs in the conformant Kubernetes service layer!


Q12: If you had to sum up your feelings about Project Pacific via a GIF, what would it be?

Tasha Drew

A12: GIF

David Stamen

A12: GIF

Dilpreet Bindra

A12: GIF


Bonus User Question

David Stamen

Bonus A: Definitely not, we utilize multi-master capability to allow services to be highly available.

Tasha Drew


A special shout out to our featured experts, Tasha Drew (@TashaDrew) and Jared Rosoff (@forjared), along with our other experts and vSphere fans from around the globe who tuned in and contributed.

Thank you joining our Project Pacific vSphere Chat, featuring some insightful (and witty) experts from the #vSphere and #ProjectPacific teams! See you next time, and keep an eye out for more information in our blog. If you would like to express your interest in the Project Pacific Beta Program you can find more information here.

Follow the additional vSphere experts who participated in this chat, including Joe Beda (@jbeda), Nikitha Suryadevara (@lafemmenikitha), David Stamen (@davidstamen) Dilpreet Bindra (@teran13) and Timothy St. Clair (@timothysc).

Remember to follow @VMwarevSphere and stay tuned for our monthly expert chats. Join the conversation by using the #vSphereChat hashtag and asking your own questions. Have a specific topic you’d like to cover? Reach out and we’ll bring the topic to our experts for consideration. For now, we can’t wait to connect with you in the Twittersphere!