This is the second of a two-part post. Part 1 is here = what is in the rear-view window? This is Part 2 = what’s out the front windshield?
This is all about changing the way the world builds software, and helping them build software that people love.
I won’t bury the lead: the next chapter for me will be at part of Pivotal.
My role is to be part of the team that is laser focused on driving the success of Pivotal Container Service (and the rest of PCF) together with the teams at VMware and Dell EMC. The job is simple - helping make our aligned Dell Technologies developer platform come together, and make our answer on the “how” (material = more important) of “digital transformation” (buzzword = less important) reach more customers.
Pivotal Container Service (PKS – the “K” is a shout out to Kubernetes) has a simple purpose – to be the best Kubernetes for the enterprise – with simple lifecycle operations (BOSH), security, multi-tenancy innovations and alignment with Google and GKE at its heart. If the purpose of Pivotal is to change the way the world builds software – PKS is the part of the platform for software that runs best in container abstractions, and joins the Pivotal Application Service (PAS) and the upcoming Pivotal Function Service (PFS) abstractions to extend the platform that is Pivotal Cloud Foundry. PKS has gone GA – and the combined Pivotal and VMware teams should be very proud! You can download it here.
Now, with that out of the way, I want to provide some transparency over how I’ve been thinking about this – purely for sharing the experience of navigating a personal change.
I knew a couple things:
- I knew personally that it was time for renewal and change.
- I knew that I needed to regain passion, and work in a smaller team – I’ve done startups and big companies, and I think it’s important to keep shifting.
- I had a passion for the CPSD business, products, strategy and people – but the opportunity latent in the business decision to simplify and align those parts within Dell EMC – it was a moment where I became free to think openly of what came next for me personally.
The moment of insight came to me as I spent a little downtime and dove into things I was curious about but didn’t have time to do previously. I curled up and started to learn a new language (Go). I fired up my GCP account, and started to play with GKE, and at the same time, cloned the GitHub repo from the one and only and @KelseyHightower: Kubernetes the Hard Way. I downloaded PCF 2.0.3 and started to play. The learning was delightful – and reminded me of my priorities.
Title means jack to me. I love roles that have low reports and high matrix influence. It was funny to read someone post that “Chad wanted to be CEO of EMC” in the comments on the El Reg article.
I would love to be a CEO one day purely from the “impact” and “accountability” part of the gig – but frankly titles really are irrelevant except their correlation (not causation!) with what I call “impact radius”. I’ve had big teams, small teams - and it doesn’t correlate with what matters to me.
What matters to me beyond my family (which always comes first) is: impact, people, learning and joy.
I want to thanks the teams at Dell Technologies (Dell EMC, VMware, and Pivotal) as I grappled with the question of what I wanted to be my next adventure.
I don’t want anyone to misunderstand this change. I truly believe that Dell Technologies is one of a few companies that can be that essential strategic partner for an enormous set of companies. There is awesome happening in every part of the stack, and each of the companies.
If you read my posts at the end of 2017 here, here and here – you can see, my passionate point of view is that customers increasingly want PLATFORMS – and you don’t BUILD a platform, you build ON a platform. The platform is something you consume, you don’t construct.
This will require businesses to of course continue to be open – but more importantly, I think players will need to have a vertically integrated, opinionated stack.
Dell Technologies clearly has an aligned technology stack – one that is emerging and we need to accelerate:
- The Developer abstraction layer using Pivotal Cloud Foundry – inclusive of app platforms in PAS, of containers in PKS, of functions in PFS, and data abstractions which runs on…
- … The software Infrastructure abstraction layer which can be any cloud, but is tightly integrated with the VMware stack inclusive of innovations that bring the SDS, SDN, SDC domains and tightly integrate them for running all sorts of workloads. This includes the traditional app stacks at the heart of most customers, AND new cloud-native workloads that expect “digital ready infrastructure” with things like NSX-T, tight integration with container/cluser managers and other abstractions – PAS/PKS MOST OF ALL. That layer in in turn runs on…
- … any hardware layer, but MOST OF ALL – is tightly integrated and available in turnkey engineered systems from Dell EMC in the form of VxRail and VxRack SDDC.
As I reflected and people gave me counsel and coaching, I realized I could build a pro/con list, and a “personal formula” for answering the question of “what is next for me”:
- It should involve an intersection of my loves – customers, partners, technology. I’m a connector, not a purist type that falls into clear “engineer” or “GTM person” definitions. I’m not Usain Bolt (amazing runner), I’m more like Ashton Eaton (decathelete). Well – more accurately, I’m a fatter, shorter, nerdier version of Ashton Eaton.
- It should move up the technology stack – stretching muscles in new ways. I want to challenge myself, force myself to do new things. I’d love to be a CEO one day – but there’s a lot to learn, fun problems to solve, and when I look in my heart, I’m more motivated by learning, impact and people than I am by titles.
- It should be at a smaller company – not that big companies are better/worse than smaller companies, rather I think it’s healthy to shift periodically (you learn different things, work different muscles – and are reminded of the strengths/weaknesses of both).
- It needs to work for my family – we are global citizens, and travel a lot, but we have a great situation living in Canada and the support structure of family to support my constant movement.
I explored internal and external options – and was very lucky to have incredible support inside and outside the Dell Technology family. For those that helped me (you know who you are!) THANK YOU.
With the time to reflect after we completed the transition – it means that I can say with total confidence – I’m facemeltingly excited about this new challenge, so excited that I want to tell you more about it!
As I said earlier, Pivotal’s purpose is to change the way the world builds software – and it’s doing exactly that at a growing list of some of the world’s largest businesses, creating some epic stories.
There are parts of this formula which have been refined over years.
I’ve been working hand in hand with many of my Pivotal brothers and sisters since Pivotal’s day one – and I’ve seen the refinement.
- Cloud Foundry was born as a very opinionated, very focused platform for cloud native apps, and this has been a huge success. But, being opinionated meant that there were very specific platform dependencies which in turn naturally create “pros” and “cons”. The “pro” was that for workloads that fit into its wonderful walled garden – customers can focus all their efforts above the value line, and move fast – knowing that every workload would pop out of the platform with all 12 factor principles enforced NATURALLY. The “con” was that at the start, the list of things that were NOT a fit for Cloud Foundry was long. With each release, progress has been made, and in PCF 2.0 – the Pivotal Application Service (PAS) is a huge leap forward – and I can say from experience that the acceleration that PCF has driven at customers of every shape and size is incredible. But still – the Pivotal Application Platform remains unabashedly clear in its purpose of being a Cloud Native app platform that is developer first, and keeps the division between app and infrastructure as sacrosanct. When a workload NEEDS infrastructure layer abstractions, needs forms of persistence, or otherwise needs a base container layer (aka doesn’t fit into buildpacks, existing tiles, whatever – then (put simply) Pivotal Application Service is not the right fit.
- Spring has become one of the most popular application frameworks and the tool of choice for modern Java, and SpringBoot one of its most popular “convention-over-configuration”, and this has been a huge success. But it’s not the only developer framework out there.
- Pivotal Labs has become one of the best, most proven ways to tackle the essence of digital transformation – which isn’t about any given tool, but changing the WAY people build software, and rediscovering the joy, the art – and tools and techniques to help them build great cloud native software. The techniques are independent of any given tool, but they do leave knowing HOW to use tools like Pivotal Cloud Foundry, and this has been a huge success. But, even as the Pivotal Labs and Pivotal way expands out in our partner ecosystem – there is a scale limit to this very intimate, very awesome way of changing the HOW of development.
If you put the 3 above together, you can see a done of awesome, but there’s something missing. The answer to that requires understanding what’s happening in “Cloud Native land”.
There is something else going on, something missing that is essential.
There is a Cloud Native world that comes “up” from infrastructure, rather than “down” from the developer – and that is the world of containers, cluster managers.
If Pivotal is to achieve its dream: changing the way the world develops software… then extending beyond the application platform it is today is needed, regardless of how well that approach works for some use cases.
If we want to support 100% of the Cloud Native workloads, by definition - Pivotal together with VMware, and Dell EMC must embrace, contribute and lead in the world of Containers and Kubernetes, along with Serverless/Functions – in addition to the “developer down” nature of the Pivotal Application Service, Spring, Pivotal Labs.
Now… If we can build a platform that embraces multiple developer abstractions (Application Platforms, Containers, Functions, Data services/surfaces and VMs when needed) – and do it in a cloud-agnostic way in a thriving ecosystem…. THEN we have a platform that can power the dream.
I recall a small meeting more than a year ago with core Pivotal leaders, VMware leaders, Google leaders where we were exploring what we could do together (it was around this timeframe where a ton of codevelopment started, KUBO was born).
I was there because alignment with Google and GKE was clearly central – but also a recognition of if we wanted to play a role of making Kubernetes simple for the enterprise – we needed to contribute new things, and solve problems for on-premises Kubernetes that Google solved because they control the full stack underneath GKE. Those problems that needed solving (and this isn’t an exhaustive list):
- The essential life-cycle capabilities: with an incredibly powerful low-level release engineering toolchain (this is tackled by BOSH), and full CI for the platform itself (this is tackled by Concourse).
- Solve things like critical networking abstractions and multi-tenancy in a generalizable way – even if we can’t control all the details of the physical network layer (this is tackled with NSX-T)
- The ability to deploy and extend the simplicity of the life-cycle down through the infrastructure layer.
The answer to part 3 is where VxRail and VxRack SDDC come into play. They represent our Dell Technologies aligned HCI platform and the base of our IaaS/PaaS/CaaS stack.
VxRail and VxRack SDDC make the virtual, software defined infrastructure move from something you construct into something you consume. They do this from the hardware to the software, all in one, all with full lifecycle management. VxRail does this inclusive of the SDC & SDS layers (vSphere, vSAN). VxRack SDDC does this inclusive of the SDC, SDS, and SDN layers (VMware Cloud Foundation – inclusive of vSphere, vSAN, NSX). You can see where this is going.
I remember thinking to myself in the dialog… this is going to be insanely fun, insanely interesting – and something I want to be part of.
PKS is amazing, and cool – even at the 1.0 release. It is an Enterprise-ready Kubernetes. It has certain things that are really cool, IMHO:
- PKS is using the pure, stable OSS Kubernetes at its heart – no proprietary extensions at this fundamental layer. We are aligned with GKE and will keep it that way. GKE uses 1.9, and so does PKS – and customers should expect that to continue.
- PKS has multi-tenancy and secure network isolation built-in, not bolted on (or just absent!). This is something so essential, and something I see consistently at customers as they look at containers inside the enterprise in prod. The devs go NUTS when they are held up because going into production requires compliance with their 12 security domains that are implemented by rigid rules, VLANs and manual firewall configurations that require a change control process involving paperwork. It can TOTALLY destroy the principles behind CI/CD and the devops operational model. Yet – this is essentially true at every single enterprise – sometimes with networking teams digging in their heels. Making NSX-T an embedded PART of PKS has been the result of awesome collaboration between VMware and Pivotal – working together to make REAL progress, REAL differentiation that solves REAL problems for customers.
- PKS has BOSH at its core – which means that deploy, patch, update of Kubernetes clusters can be totally automated. This isn’t just about lifecycle – it’s about essential platform security. It also means that for the ecosystem that is using BOSH for automation of their own tools – we can make PKS an easy target. There’s a huge opportunity there for Redis, Mongo, Cassandra, Postgress, Greenplum and the whole data fabric ecosystem.
If take PKS (or PCF 2.0 as a whole) and run it on a VxRail – that means that the WHOLE STACK is automated – with only one exception (the physical network firmware layer). Day 2 lifecycle is no longer your problem – all the way up (including the NSX-T components). We would have a Pivotal Ready Architecture – a clear, opinionated “how” of the on-premises technology part of Cloud Native and digital transformation.
Note one roadmap “discontinuity”: Today, VMware Cloud Foundation (and therefore VxRack SDDC) don’t support NSX-T today running on ESX hosts (only NSX-V). Furthermore there aren’t yet simple automated workload domain deployment models for multi-availability zone PCF deployments. Those are why today VxRail is the “easy button”
Trust me – we’re on it as a team. You can imagine that at a point, we will make starting with a VxRail easy, upgrading to a VxRack SDDC simple – and both can be fully lifecycled platforms underneath PCF – inclusive of the Application abstraction (PAS), the container abstraction (PKS), the function abstraction (PFS) – and perhaps even we will include data abstractions (from object stores to all sorts of data fabric layers). Both also expose the VM abstraction when you need it.
People debate the pros/cons of every part of the stack – but frankly, I think that’s the wrong topic to get into bar-fights over. If you could have a turnkey platform on-premises – and it was multi-cloud, bridging transparently to AWS with VMware Managed Cloud when you need an IaaS, bridging transparently to GKE, or frankly to any cloud under the PCF layer – that would be incredibly powerful.
So – as you can see, I’m really excited about it!
What am I going to be doing specifically?
- I’m going to be in a glue role, a bridge role – I love these roles, and it’s my nature to be a “connector”.
- I will make sure that Pivotal, VMware, and Dell EMC work hand in hand from how we work in the market, and how we bend the roadmaps to be aligned – and deliver on the dream for hundreds and then thousands, and then tens of thousands of customers and partners.
- I will ensure the execution and scaling of the PKS effort amazing people like Colin Humphreys, Paul Fazzone, Craig Steel, Scott Yara, James Waters, Ian Andrews, Raghvender Arni, and many, many others have been working on so hard and should be so proud of.
- I will make sure that my brothers and sisters at VMware and Dell EMC working on the aligned HCI stacks of VxRail and VxRack SDDC using VMware Cloud Foundation are in lockstep, and THE best way to deploy Pivotal Cloud Foundry - in each of its 3 components (singularly or together): PAS, PKS, and PFS.
- I will take a little time to be a full time dad in the month of March, and start in full force in April.
I’m joining a team I’ve felt like I’ve been part of for years – but energized in a way that is new.
I’m excited, I’m renewed – and I’m in the never-ending process of redefining and rediscovering myself.