The year is 2020 and you are responsible for ensuring the efficient and reliable operations of millions of dollars of cloud computing infrastructure. Things have changed a lot in the past few years, even more so from the days when you first moved from on-premise and into AWS.
Today your business operations and cloud infrastructure are tightly mapped to customer demand and usage. You don’t pay for idle compute time and your engineering teams are always looking for ways to improve a product’s operating cost by optimizing your compute, architecture and data pipelines. In fact, your engineering team has helped the business boost margins by over 50% this year, without any price increase, by optimizing system efficiency for the greatest revenue generating customer transactions.
But it wasn’t always this way. It wasn’t until the advent of Serverless cloud technologies and the rise of FinDevOps practices that the business was finally able to align capital flows with technology decisions.
Serverless: Addressing the Technical Challenge
At some point in our future, this term serverless will disappear. We will just call it “cloud”, but today is not that day. Because the serverless revolution in cloud computing is so significant, it is necessary, for this moment, that we call it something different. Serverless unlocks our ability to tightly fit the usage curve of computing power with the actions and needs of customers. It fundamentally changes how engineers will approach the design of systems and software and we need to call it out.
Properties of Serverless Systems
Something so significant has a deceptively simple set of properties that took decades of engineering effort to achieve.
- Event driven
- Invisible infrastructure
- Automatically scales with usage
- Fault tolerance and high availability built in
- Never pay for idle (except for storage)
Functions as a Service (FaaS) offerings like AWS Lambda may have kicked off the serverless revolution in 2014, but the concept of serverless is actually rooted in the very first cloud services ever launched, AWS SQS, and S3. Both of which check off all of the properties, perhaps even more so, than AWS Lambda.
Serverless Isn’t Just FaaS: Serverless is a Spectrum
Serverless encompasses a broad spectrum of services today, well beyond just FaaS. In fact, most organizations today have some serverless in their environment, only they call it by another name, Platform as a Service or PaaS. By mapping your PaaS usage against the serverless properties you can determine where you fall on the spectrum and determine just how serverless your systems are. The further right on the spectrum the more serverless the resource and the greater breadth – and depth – of information available about your use of that resource. More importantly, the further right you are on the spectrum, the more prepared you are to push forward the cultural evolution of DevOps to embrace and address the financial impact of software on your business.
FinDevOps: Addressing the Cultural Challenge
While serverless addresses the technological limitations, the culture of DevOps needs to evolve to treat cost as a first-class operational metric. With finance and DevOps aligned, the business is now in the position to make informed technology choices that are truly aligned with business goals.
Simon Wardley first kicked off this idea of a new operational practice called FinDev in his blog post Why the fuss about serverless? and even more recently in a tweet, Simon doubled down on his statement that serverless would fundamentally change how business and technology work together.
The way we monitor applications will shift towards capital flow within applications. Refactoring will have financial value. The way we invest, the business models we create … all is about to change. The next decade will be thrilling. And you thought cloud / DevOps was big. Lol.
— Simon Wardley (@swardley) April 19, 2018
Simon has made this point repeatedly, and we couldn’t agree more (some of which I discuss in more detail in my recent talk Site Reliability in the Serverless Age). With the advent of serverless and granular usage based billing, organizations are now able to quantify the financial benefit of investing in coding or system improvements, just as they do with other parts of the business that are more visible and tangible.
Addressing this however doesn’t come from technology alone. Every disruptive evolution gives rise to a new culture as well. DevOps has to evolve to see capital flows as critical metrics throughout the lifecycle of software development and operations. We call this emerging culture FinDevOps in recognition that just as operations and development merged to create a new culture around Cloud technology, finance must also be integrated if we are going to, as Simon says, fundamentally change how we build businesses around technology, and how we code.
Cost as a First Class Operational Metric
The business and technology teams of the future will be much more aligned than they are today. Through a combination of Serverless technology and FinDevOps culture, teams will accurately know the operating cost for every product, every feature and every workflow. No more guessing. Development, operations and finance all working together seamlessly in the best interest of the business. Businesses that embrace this vision will be the businesses leading in the new world.
If you are interested in learning more about how CloudZero can help your business make this transformation and embrace FinDevOps, contact us here to learn more about our platform and services today.