I’ve been spending a bit of time trying to distil the large amount of information coming out of VMware of late around modern applications and the Tanzu portfolio. I’m happy to to admit that my background has always been in the infrastructure side of things, and historically that’s where VMware as a company sat; focussed on the platform which happened to host applications. The world has and is changing though, driven by public clouds and the next level of delivering speed, agility and self-service is by increasingly partnering directly with the end users of the cloud service, whether that be public or private cloud.
As such, understanding this new reality is key to architecting, engineering and managing modern cloud solutions, including the different customers and their diverse needs. Virtual machines aren’t going to go away but the use of ‘cloud native’ application architectures and Kubernetes (which abstracts the infrastructure layer) is increasingly part of a hybrid landscape, which has to navigate a tricky path between freedom for developers to innovate at speed while allowing administrators to maintain the stability which has always been a pillar of vSphere. This is where Tanzu comes in. I thought this information might be useful to like-minded folk who no doubt have the same questions I did. I’ve included a good few links with supplementary information should you wish to dive deeper.
So what is Tanzu? Tanzu is an umbrella under which VMware’s modern application or container focussed products reside, much is the same way as vRealize covers the cloud management products such as vRealize Automation and vRealize Operations. Tanzu has a number of different layers.
TKG
Starting at the bottom is Tanzu Kubernetes Grid (TKG) which is VMware’s certified Kubernetes distribution. TKG is what is actually embedded into the hypervisor in vSphere 7 to run Kuberenetes and containers natively (what is now known as vSphere 7 with Kubernetes and was perviously referred to as Project Pacific). TKG has a number of different editions as outlined below:
The Solution Overviews for TKG, TKG+ and TKGI are a great reference point for further information, but the main differences are:
- TKG+ includes CRE support – essentially this means proactive support and guidance rather than the reactive break/fix support through GSS with the basic TKG
- TKG+ supports VMware Cloud on AWS
There are other differences which are articulated in KB 78173. TKG and TKG+ are available from vSphere 6.7 U3 in a standalone architecture. However, for full native integration (Project Pacific), not only is vSphere 7 is required, but it must be delivered through VMware Cloud Foundation 4, which also includes NSX-T 3.0 and lifecycle management capabilities to make this properly hang together. See here for more information on this. The difference in these architectures is nicely shown in the diagram below (taken from this blog).
Note that at the time of writing, TKGI is announced but not yet generally available.
TAS
The Tanzu Application Service (TAS) is an application platform that has evolved from what was Pivotal Cloud Foundry, then Pivotal Application Service and latterly Pivotal Platform. TAS provides an additional layer of abstraction which aims to allow developers to be more efficient by focussing on applications and code rather than the containers across different environments (dev/test/prod) and the orchestration around that. One of the ways it does this is through buildpacks.
TAS also supports multiple clouds. Many customers use TAS to provision to native AWS EC2 and there are a number of reference architectures for differing cloud platforms, both private and public.
TAS comes in two editions – ‘for VMs’ and ‘for Kubernetes’ (in public beta). The difference between these two is TAS for VMs is uses Diego for container orchestration while TAS for Kubernetes is rearchitected to be more deeply integrated with Kubernetes and uses Kuberenetes containers for the control plane.
TMC
Tanzu Mission Control (TMC) is an overarching management platform (a single pane of glass, if you like) to manage Kubernetes clusters across clouds. TMC can provision and manage TKG instances and furthermore, you are able to attach any CNCF-certified (Cloud Native Computing Foundation) cluster, which includes the likes of Google’s Kubernetes Engine (GKE).
VMware Hands-On Labs is an excellent resource to get exposure of TMC. HOL-2032-01-CNA goes through the process of attaching and provisioning clusters to AWS and I’d highly recommend giving it a look.
Observability by Wavefront
Tanzu Observability aims to provide enterprise-grade observability and analytics to be able to pro-actively monitor and measure this at an enterprise level. It does this through ingesting metrics from applications, allowing visualisation and correlation to more easily gain insights into issues. As you’d expect from a tool of this nature it allows alerting to be tuned as well as supporting automatic remediation.
CloudHealth
While Tanzu Observability provides insight into metrics, CloudHealth gives insight into costs and resource utilisation across these different clouds. It also details trends and the ability to right-size resources which is crucial to keeping costs in check and avoiding sprawl.
How it all hangs together
Below is how all of this hangs together, TKG is VMware’s Kubernetes runtime which can be deployed on-prem or in VMware Cloud on AWS. Tanzu Application Service can abstract the infrastructure, whether on-prem or in public cloud (e.g. AWS EC2) and Tanzu Mission Control, Observability and CloudHealth provides insight across this multi-cloud landscape.
vCloudburst Blog 
Pingback: Tanzu Kubernetes Grid Integrated GA | vCloudburst Blog