We are in a cloud native talent crisis. It is not going to end soon.
Given the depth of the crisis, it is pretty amazing to consider how quickly the cloud operating model has grown – from the public cloud to the private cloud and even out to the edge.
Our clients and partners are all angling for more cloud talent. It is the primary skill we seek in new employees. Even accounting for the core principles of containerization, orchestration, automation and APIs, you are still left with a small pool. Layer on cloud-specific expertise for one of the Big Three – much less any of the longtail providers – and it gets pretty bleak.
It doesn’t have to be this way, however. You can grow your cloud operating model talent and expertise through a disciplined selection process. What this means is that by selecting for simplicity and multi-cloud capabilities in core areas, an enterprise can position itself for success with the talent it already has.
This will not and should not stop the cloud-native talent march, but it can dramatically alleviate the pain and position the organization for future success. This will require choice, there may even be tradeoffs, but the approach will deliver scale that is independent of human resources.
Here is a checklist of items to consider as you build and refine your talent-constrained cloud software stack:
- Choose software. It may seem obvious, given that the cloud is a software-defined world, but there are a lot, and I mean a lot, of appliance-slinging salespeople out there trying to convince you that they can deliver cloud-like simplicity. They can’t. You can’t containerize an appliance. Appliances may run in the cloud, but that doesn’t make them cloud native – it just allows them to hang on for a little bit longer.
If you want to address the cloud talent crisis, don’t do it by buying more hardware appliances – it is literally the worst thing you can do.
- Choose software you can try. That means there is a download button somewhere on the website. We buy software all the time in the marketing function, from RFP response software to ad-serving software. If we can’t take it for a spin, we can’t assess the usability. Demos don’t tell you that, experience does.
- Optimize for cloud native standards and/or open source projects with large communities. In the storage world, the standard is S3. Full stop. Everything else is legacy, bolt-on nonsense. In the compute world, the standard is Kubernetes. Full stop. The list goes on – containerization over VMs, GO over Java.
Choose technologies that use the standards. This does not require making exotic bets at this stage. Service mesh isn’t required by most organizations. Don’t sweat it. If you get the building blocks correct, service mesh will be an option.
Because when you optimize for standards, you inherently get integrations. MinIO has thousands of such integrations, most with companies we have never spoken to, don’t have agreements with and may never officially interact with? How? Because we are built on the standard principles of the cloud and have a download button. Cloud native talent recognizes cloud native talent so you don’t have to have a corner on the cloud native talent market.
Select your way to success. Pick standards-based software that comes with integrations built in. Want evidence? VMware picked a 40-person company in MinIO to be the design partner for its biggest product launch since the VM. Why MinIO for Tanzu? Because they knew we did Kubernetes-based object storage, were software-defined and would bring thousands of applications to Tanzu on day one. If they ran on MinIO and MinIO ran on Tanzu – volia, instant credibility.
Select your way to success.
- Optimize for what is important to the enterprise. Again, this seems simple, but we often encounter decisions that were made before MinIO that have constrained options in some area of real importance. If security is important, ensure your selections enable you to bring your own IAM. If performance is important, ensure that your selections don’t have some underlying issue that limits performance (small file chattiness over the network or old, repurposed hardware). There are things that matter to the organization – make sure they are addressed in the selection process and done so in a cloud-native way.
- Graphical User Interfaces. The cloud-native world grew out of CLI, Helm Charts and YAML files. It is intimidating. It is unyielding. This is why high level services in the public cloud did so well. They abstracted complexity. What Snowflake, Dremio and Starburst do is inherently very complex. They have all worked tirelessly to abstract the complexity associated with analytics, data and storage. This in turn has massively expanded their audience.
They did it through graphical user interfaces. Can you still drive those products from the CLI? Yes. But that level of talent is what is so scarce. Graphical user interfaces are the key to scale.
Expand the audience. A quick example for you to try. Here a user will set up multi-tenant, object storage as a service with less than 20 clicks. Frankly, you can do this in less than five, but we wanted to show all the options. The point here is that you can productize devops to the extent that anyone, including old school IT admins, can drive and manage massive, cloud-native infrastructure. Again, this isn’t an appliance interface. This is creating and deploying cloud-native storage infrastructure on commodity hardware. Theoretically, this should be hard. Click through the demo and let us know how hard that felt.
6. Choose simple everywhere, but don’t let simple be a proxy for under featured. By selecting for simple, the enterprise creates more entry points to the cloud operating model. More entry points and more employees using cloud technologies should not be concerning. Rather, if you select for simple, but don’t yield on features, you will end up with a solution that is exceptionally resilient to human error without frustrating users in the process. Actively select for this and you will uplevel the cloud-native talent in your organization without them even knowing it.
This checklist is highly achievable. It requires discipline. It requires effort. But it is within reach for the vast majority of enterprises - particularly those with one foot in the cloud already.
The talent gap isn’t going away. Cloud and data leaders need to acknowledge that reality and start to build a foundation that gets the majority of their organization to the cloud operating model today while keeping optionality for tomorrow.