Whether you examine large webscale companies such as Apple or small startups, there’s a consistent theme regarding the leveraging of Public Cloud. That theme is that Cloud has a place for specific business and technical drivers. One of the original thoughts regarding Cloud is the concept that the majority of organizations would outsource their infrastructures to Public Cloud services such as Google Compute Platform (GCP).
Spotify is an example of a company that decided to bet big on Google Cloud. Spotify decided to leverage GCP to power its analytics and infrastructure. On the other end of the spectrum, you have companies such as Apple that have invested billions of dollars in physical infrastructure. Other headlines show businesses of all sizes either moving to the Cloud or bringing workloads back from the Public Cloud. So what gives? What are the drivers behind the decision to use the Public Cloud and when to stick to private infrastructure? Let’s take a look at both the business and technology drivers.
While I expect to talk about cost from a high-level, it’s not the primary driver for moving to the Public Cloud. The Cloud conversation begins at the speed of execution. There’s a reason why there’s so much hype around Public Cloud providers. Simply put, Public Cloud provides the capability to execute ideas faster than legacy infrastructure. The initial conversation isn’t about capital expenses vs. operational costs. It’s not even about cost avoidance. The first discussion focuses on the creation of a strategic business advantage.
Traditional industries such as transportation, financial services, and retail undergo disruption by relatively new players to their respective markets. These new players transform the way goods and services are purchased and delivered. Taking a cursory look at these players you’ll notice that transformation starts with leveraging what sounds like unlimited computing power. How exactly is a retail startup able to compete globally, conduct and process a massive number of transactions daily and pivot based on the analytic data collected from the operations? The technical cost is an incredible barrier to entry into this global market. The time needed to build the infrastructure is just as daunting.
Enter the Public Cloud. New players quickly build scalable platforms that are torn up and down as the business grows and pivots. The infrastructure is an asset vs. a barrier to entry. A garage startup using a credit card and some great code can virtually sell parking spaces through the entire U.S. Incumbent competitors need an answer to these agile competitors. A deployment time of 9 months for new infrastructure results in continued broad market share losses. Public Cloud provides a level playing field in fast-changing industries.
Not all industries are fast changing. Some markets have dampeners on them that prevents fluidity. These are areas where fast isn’t always a good thing. Think Pharma, Utilities, and some government functions. There are specialized use cases for Public Cloud in each category, but on the whole, the industries are yet to be disrupted by startups leveraging Public Cloud infrastructures.
Private vs. Public
A simple question is why not build the elasticity internally? Indeed, this is the route Apple is taking with building physical data centers and presumably moving workloads away from the Public Cloud. However, every company doesn’t have the financial or technical resources of Apple. Most organizations still struggle with creating self-service portals and chargeback models for their virtualization environments. Companies such as Google have dedicated entire product teams to building multi-tenant/multi-nation infrastructures that have proven web-scale performance and reliability. Google provides a zero-trust network infrastructure as a default design. Most organizations struggle to deploy micro-segmentation in their data centers without breaking existing applications.
Private Cloud has valid use cases, however, if the aim is to solve the agility challenge then undertaking a Private Cloud design and build is a project best approached with care. The very opposite results may occur if the attempt is to reduce the time to implementation if you become bogged down with years of trying to replicate Public Cloud services internally.
Now we get to the cost portion of the conversation. There’s little doubt that the startup cost for Public Cloud is less than building a dedicated infrastructure. Startups are not the only companies that benefit from the lower implementation costs of Public Cloud. In my call out of industries such as Pharma, Government, and Utilities, each industry can take advantage of Public Cloud cost for several projects and workloads.
Public Cloud allows for the testing of new ideas regardless of the industry. If a group of scientists or engineers want to test a new product or run analytics, the pay for what you use model of Public Cloud has great appeal. A product or project ideas are quickly spun up and torn down in Public Cloud. There’s no long procurement cycle or depreciation cost to worry about if the project fails. If the project succeeds and demand creates the need for additional capacity, Public Cloud easily grows to meet demand.
Like any other investment, organizations revisit the cost of Public Cloud services. For certain workloads, there’s sometimes a point when dedicated infrastructure is cheaper and more convenient than Public Cloud. The poor cost comparison for Public Cloud happens when workload needs stagnant or become predictable. What is true at the onset of a project isn’t necessarily true a few months or years later.
Public Cloud is a great option when agility and cost flexibility is needed. Mature or regulated industries should carefully consider when Public Cloud provides a tactical advantage for specific workloads and projects. There isn’t a single set of companies that should consider Public Cloud. At this point there’s a use case for Public Cloud for most organizations.