Reducing the Cost of Failure: The Cloud's Secret to Hyper-Growth
By Dave Kuhn, Group Director, Technology, Growthops
Dave Kuhn, Group Director, Technology, Growthops
At the Sydney Google Cloud Summit back in September it was noted that only 10 percent of computer workloads globally had migrated to the cloud. Just 10 percent! Big iron still rules for now it seems. As someone who’s been working with the cloud for over a decade now that figure came as a shock. Cloud tooling, technology, processes and practices have matured fantastically over the last few years and it can be easy to think that we’re well past the point where ‘going cloud’ can move the needle.
This statement had me reflecting on any number of conversations we’ve been having with companies recently on growth where a consistent theme has been emerging. While publicly, everyone loves to talk about disruption, new markets and name dropping the latest technologies they’re investing in, behind closed doors it’s a very different story. Moving from prototype to production is fraught with challenges that largely tie back to their considerable on-premises infrastructure footprint.
Why should on-prem infrastructure have such an effect on growth? It’s instructive to look at major disruptors such as Google, Salesforce, Amazon, Facebook, Uber, AirBnB, Netflix, Dropbox—the list goes on. While their missions, products and markets differ, one thing they share, almost without exception, is they were either born in the cloud or outright created their own. This isn’t to discount or downplay the role of culture, leadership, hiring, structure and economic factors, but it’s interesting to consider nonetheless. Why do they all share this common thread?
The reason is simple: to innovate at pace you need to substantially reduce the cost of failure.
A choice between slow failure or learning fast
Practically this means severing the link between procurement, IT operations and security—which necessarily operate on different time scales—and product development. Unless these ever move in lockstep you will not be able to achieve hyper-growth. Cloud has enabled these companies to move at two speeds—provisioning rapidly without long term commitment, while managing costs, achieving operational excellence and staying secure.
Meeting with a client recently I witnessed a heated debate between a Head of Development and the Head of IT Ops. They were arguing about the amount of time it took to provision a new production server into their “private cloud.” IT Ops said it was two weeks, Development strenuously insisted the actual timeframe was closer to two months.
Cloud has enabled companies to move at two speeds—provisioning rapidly without long term commitment, while managing costs, achieving operational excellence and staying secure
Building new products and services requires experimentation. Predicting demand is uncertain, at best. Who knows if you’re going to be Facebook or Google+? Even the biggest companies get it wrong, why will you be any dif ferent?
Traditional procurement demands that up-front choices be made about server size and storage capacity, which translates into an investment that then takes years to achieve a return on. Worse still, what if you get it wrong and under provision? The ABS learned this the hard way in the 2016 census. These are low value and time-consuming decisions to be making which carry significant financial consequences. Worse still, it adds nothing to your understanding of the problem or the customer.
Likewise, the lack of pervasive automation around internal infrastructure leads to constant reinvention of the wheel every time a new service is to be deployed. Discussions around network topology, firewall rules, configuration management and security standards dominate, taking away valuable time and resource from solving hard problems and building amazing experiences. Meanwhile, sophisticated cloud native businesses leverage extensive automation to deploy services—even construct whole new environments—within minutes, using standardised pipelines with in-built security and quality assurances.
Of course, even for those convinced of the link between the cloud and growth there are substantial challenges and serious objections to overcome. Privacy and security feature prominently along with concerns around billing and vendor lock-in. The good news is that cloud vendors have invested heavily in recent years in industry recognised certifications such as ISO, SOC, HIPAA and IRAP. Moreover, they’re constantly expanding their data centre footprint into new jurisdictions addressing many concerns around data sovereignty. Consequently, it’s becoming increasingly difficult to argue the case that internal infrastructure offers anything like the same level of rigour and safety as is found in the cloud.
The three Rs
No matter where you’re at in your cloud migration journey I find it helpful to break things down into three Rs: Ringfence; Refactor; and Rewrite.
Ring-fencing involves identifying which legacy apps can’t move owing to unusual hardware (think mainframes, AS400s etc) and forbidding anything but essential development (e.g. critical bug fixes). Connect them to the cloud via an interconnect and wrap them (cloud-side) in a modern facade to allow other apps and services to integrate.
Next, identify which applications can be refactored into containers (third-party and large legacy apps) and deployed to an orchestration service like Kubernetes.
Lastly, kick things into high gear by rewriting fast moving apps using serverless technologies like Google Cloud Functions or AWS Lambda.
While it may be unsexy, for many of our clients, moving to the cloud is the highest value thing they can do today to unlock growth within their organisation. Businesses that are committed to growth and innovation are quietly getting on with the job of steadily moving to the cloud. With such a small percentage having made the transition thus far, it’s fair to say that there’s still a great deal of opportunity to join a small cadre of cloud enabled businesses geared for hyper-growth.
