You don’t have to be a Wall Street professional to understand the reasons to invest in more than one company. So, why would a business choose a strategy of investing in a single cloud? It doesn’t make sense in today’s marketplace.
Initially, cloud providers like Microsoft Azure, AWS or Google Cloud were commodities using pricing as a competitive advantage. But today, these providers are differentiating themselves with unique value-add services and features that make sense for different businesses. Google, for example, offers Vision AI services, which for retail business will enable them using facial recognition to connect online and offline identifies for loyalty programs. But if a company is focused on replatforming some of its legacy applications and bringing them to the cloud, it will probably want to use Microsoft Azure SQL and Azure Active Directory.
In addition to features, companies want to consider their needs for locality, regulatory or service-level agreement challenges that each provider might address differently. When going multi-cloud, they should take advantage of “Best of Cloud” as well, leveraging the optimal feature sets from each provider that makes sense for the business.
Ultimately, if a company is not moving toward building a cloud ecosystem of microservices and applications, then it will likely fall behind competitors. Businesses should consider these four principles for building a best-in-class multi-cloud strategy.
Understand the Accordion Thesis
IT is like an accordion and at any moment, there are parts of it that are expanding and parts of it that are contracting. The contraction is standardisation, where all diverse choices are gradually squeezed down to a single standard (like the 19-inch server rack); whereas, the expanding parts are the innovations, such as machine learning, chatbots and AI. On the innovation side, an organisation needs to be able to select and experiment with an infinite number of choices. The trick in managing IT is understanding standardising and where you’re innovating, and to try not to standardise on the innovation side or innovate on the standardisation side.
The hardest part for most IT organisations is the “Fat Middle,” where it’s unreasonable to try and force a single standard choice, but where a plethora of choices is just unnecessary overhead. Most healthy businesses can tolerate two to six choices within each category, for example having both SQL Server and PostgreSQL available.
Companies can use the Accordion Thesis to inform “Build versus Buy” decisions as well. It can also be helpful for avoiding one of the greatest business risks: analysis paralysis. Do pilots, learn things, but have a bias towards building cloud native applications or microservices that will facilitate innovation, and leaning on standardised services provided by the major cloud leaders.
Take control of what’s already in the cloud
Every Fortune 500 company is using the public cloud, but most likely are not using it strategically or even on purpose. Much of this use is rogue lines of business, such as applications that the marketing team utilised because they wanted to move fast without informing the rest of the organisation. A part of a company’s cloud strategy is taking control of the ways it’s already using the cloud. As the number of applications continues to expand geometrically, businesses can find it harder and harder to understand the operational footprint of this sort of tactical ‘mix and match’ approach.
Build your own ecosystem, don’t depend on others
The more businesses have built against one cloud provider, the more likely they’ve inadvertently embraced a lot of -isms of that provider around things like storage, authentication layer, and access or network topologies that are not portable when they shift to a multi-cloud approach. The sooner a business adopts a multi-cloud strategy, the less it’ll have to rework to make it possible.
Once a business has identified the differentiated features and services it wants to take advantage of, it can bind them to its applications using a system broker. Imagine a service broker like a fund manager: it has access to all the different options and connects them together. This ecosystem is what will differentiate innovation, enabling companies to create additional applications that take advantage of best-in-class features from leading cloud providers.
Don’t conflate the present with the future
Another reason to diversify a cloud portfolio is that the future is unknown. In 2000, people thought the space was so crowded that “it was finished” when up against leaders like Altavista, Lycos, Yahoo! and Excite. Today, there is a similar — albeit misguided — implicit assumption: “Cloud is done because the major players are here.”
Just because AWS was there first, and has tapped into a decade of pent-up demand, it’s foolhardy to believe AWS will always be the market leader. When a business adopts a standard that has at least two vendors, then it can protect itself to some degree against the unknown.
Of course, the engineering team may have some resistance to a multi-cloud approach because they assume there is upfront work to learn a new cloud’s commands. This may seem obvious to some, but the languages of Microsoft Azure, AWS, or Google Cloud are like romance languages: they’re all very similarly rooted. There are minor changes for each provider, so this shouldn’t be a major hurdle, or, a business can use software which automatically translates between these cloud infrastructure providers.
There’s a reason that the top banks, auto-manufactures, telecoms and other Fortune 100 companies are all adopting multi-cloud strategies: innovation is key to their success. The speed of innovation is now more paramount than ever, especially given that competitors have a lower barrier for entry with the reduced cost of outsourcing infrastructure.
Interested in hearing industry leaders discuss subjects like this and sharing their experiences and use-cases? Attend the Cyber Security & Cloud Expo World Series with upcoming events in Silicon Valley, London and Amsterdam to learn more.
