I’d say that the biggest challenges are operability and spiraling cloud costs.
The number one thing that you hear people saying today – or complaining about today – is their cloud costs. But the real problem isn’t their cloud cost – it’s not the expense itself – it’s that they don’t really understand the value they’re getting out of cloud. They can’t connect their cloud spending to the business outcomes that they want.
If you go back before cloud, when a business wanted to launch a new product or service, they’d figure out the size of the market, they'd figure out the size they could get, they’d figure out what their software and hardware costs were, and then they were able to determine whether it was a good decision to even create this and know when they’d made a profit – or stop and give up on it. Now we’ve hidden all that from them, and basically we’ve left the business behind.
When we talk about operability, I think it’s important to look at how we got here. The first wave of cloud platforms, I like to call the walled garden. Walled gardens are very standardized and extremely easy to operate. But they’re not very flexible and developers can’t just develop whatever they want in whatever language they want. So they’re restrictive.
The next phase in cloud platforms was a reaction to that inflexibility. That’s where you saw Kubernetes, and especially DIY Kubernetes, being built. And those systems were incredibly flexible. But they were extremely difficult to operate because there’s no standardization and they were very difficult to automate.
It all came to a head last year when there was a security vulnerability in a library called Log4j. The companies that built these flexible platforms weren’t able to adjust and fix this in a reasonable time frame.
Everyone knows they have a problem, they just don’t know how to fix it. And that’s really where we’re going with the third wave of cloud platforms.
The solution to cloud’s operational, flexibility and cost challenges is the third wave of cloud platforms and that is declarative, autonomous cloud platforms.
In a declarative, autonomous cloud, both developers, business, finance architects can declare what they want to happen. And the system just makes it so. So it gives both the flexibility and the operability.
To take that even a step further, a business should be able to declare what they want the cost to be. If they want this to be one cent a transaction, they should declare that and the system should try to make that happen. And if it can’t, they can decide to degrade performance.
Cloud FinOps is an evolving financial management discipline where engineering, technology, finance and business all collaborate to control and understand cloud costs.
A declarative, autonomous cloud is just the start of what a FinOps model should be. There’s also organizational and cultural changes that are a part of that.
I like to look at it as similar to when we all moved to agile development. We changed how we were organized so that we could develop more modern software. Well, FinOps does about the same thing. We change how we’re organized and how we operate so that we can understand and control cloud costs.
Organizational changes are always difficult. Everyone wants an easy button. They’d like to buy a simple tool and have that fix everything. But a simple tool can’t make the changes necessary to fix this problem.
Adopting the FinOps model requires trust between the technologists and the business. And that takes time to develop. UST is adopting the FinOps foundation model and we’re working with industry experts to build strong customer relationships and help them on the next stage of their journey to cloud.
The reason everyone built all these DIY Kubernetes solutions was because they thought that that flexibility would make the developers more productive. But it didn’t.
DIY systems are often more complex. And they actually increase the time from ideation to the first line of code, which is a measure we use.
Companies have also built their cloud platforms for 2% of their developers, the squeaky wheels, the rock stars that come and say that they need access to Kubernetes internals, they need access to every function on every cloud.
But in reality, 98% of an enterprise’s developers want to write their code, get their bonus and go home and spend time with their families. That’s what they really want.
The answer is the third wave of cloud, which is both operable and flexible through declarative languages and allows developers to quickly declare what they want and have that developed in the pipeline.
contact us