“When it comes to cloud providers, an open relationship is preferable…”
The “Battle of the Clouds” was an important campaign in the American Revolutionary war, in which George Washington was forced to retreat due to heavy rain and mud, writes Aron Brand, CTO, CTERA. 243 years later, there’s another battle of the clouds going on, and this time the combatants – Amazon, Microsoft and Alibaba – are not showing any signs of giving up.
Amazon Web Services (AWS) is still the unrivaled leader for cloud infrastructure services with a gargantuan 47% market share in 2018 (Gartner). However, this lead has been eroding steadily since 2016 as Microsoft and Alibaba are eating Amazon’s lunch with their hyper growth.
Despite having triple the market share, Amazon is very concerned about Microsoft Azure’s rapid growth, fueled primarily by a wave of relatively late adopters migrating their companies’ IT systems to the cloud. As these risk-averse organizations and departments are heavily dependent on Windows servers, they tend to consider Azure to be the safest choice based on its integration with existing Microsoft software.
Microsoft cemented its status as a major nuisance for AWS last year when it narrowly beat the cloud behemoth head-to-head for a $10+ billion Pentagon cloud-computing contract – a win that Amazon is hotly contesting in court.
Somewhat surprising are the tenuous positions of Google and Big Blue. Despite its technological prowess and superb cloud infrastructure, the Google Cloud Platform has shown slow growth over the past few years compared to Microsoft and Alibaba, and is struggling to capture the enterprise market. IBM, a recognized enterprise IT leader, has not been able to leverage this base when it comes to selling cloud infrastructure, with its market share flatlining since 2016.
Amazon’s Siren Song of Proprietary Cloud Services
So how does this market consolidation impact enterprises’ cloud strategy? To answer this question, we first need to understand how market dynamics affect cloud providers’ business behavior. As the undisputed market leader, Amazon has a vested interest in pushing its huge wealth of proprietary cloud offerings to bind its customers, while keeping smaller rivals at bay. I often attend Re:Invent conferences, where Amazon has made it a habit to wow its users with a plethora of new services. At the same time, the term “multi-cloud” is understandably absent from their vocabulary.
In stark contrast to AWS, challengers like Google are trying to win enterprises’ affection by investing in open cloud standards that encourage a multi-cloud strategy. The best example of this is Kubernetes, Google’s open source container management tool, which is becoming a de-facto standard for workloads that run in multi-cloud environments.
Embracing the Multi-Cloud Mindset
The consolidation of the cloud services market underscores the importance of having a coherent multi-cloud strategy. AWS and Azure already account for well over 60% of the public cloud market, and growth trends suggest that this dominance will continue in the coming years. Accordingly, as an IT executive – if you’re not doing so already – it’s extremely likely that in the next couple of years you’ll want to use AWS and Azure services in parallel (with or without other clouds).
In the short term, if you decide to run keep workloads on a single cloud vendor, prefer cloud-agnostic solutions wherever possible. For example, if you are an Amazon customer, prefer Amazon EKS Managed Kubernetes Service over the more proprietary ECS. For Infrastructure as Code, I’d prefer Terraform over Amazon’s proprietary Cloud Formation.
If budget allows, I’d also recommend using AWS as your primary datacenter, and Azure as your secondary DR datacenter (or vice versa), with your data being continuously replicated. This is somewhat more expensive than using a pair of availability zones of a single cloud provider, but doing so would guarantee that you can painlessly switch cloud providers in the future.
How a Multi-Cloud Strategy Supports Your Long-Term Business Goals
Here are three ways that multi-cloud state of mind helps enterprises achieve business goals:
- Bargaining Power – In order to compete with the market leaders, smaller cloud providers may offer the same level of service at a lower price. If you are tied to a single cloud provider, you don’t have the leverage to negotiate better pricing from your existing vendor.
- Business Agility – A multi-cloud ready architecture makes it easy to switch to another a cloud provider to enable certain workloads to benefit from new capabilities not available in your primary cloud.
- Hedge Your Bets – Moving away from an on-prem datacenter to the cloud is good insurance against hardware obsolescence. But if you’re relying on proprietary API’s of a single cloud vendor, you’re risking another, potentially worse form of obsolescence. By relying on open standards and solutions that are supported by multiple vendors, you’re much less likely to need to change your systems due to end-of-life services or API changes. This prolongs the useful life of your systems.
Don’t Put All Your Eggs in a Single Cloud
Even if you’re deeply enamored with your existing cloud provider (and there’s much to love in the ones mentioned above) and you’re keeping it exclusive for now, it’s almost certain your company is going to want to go multi-cloud in future.
Accordingly, it’s important to plan ahead and resist the temptation to overly rely on the latest and greatest proprietary services and APIs. The more you do, the deeper you sink into a proprietary ecosystem.
When it comes to cloud providers, an open relationship is preferable to monogamy. As a decision maker, whenever possible, choose services and products that are open and cloud agnostic. By demanding multi-cloud capabilities from your software vendors, you can shift the power dynamic in your favor, enhance your bargaining position and remove the barriers to cloud migrations. Multi-cloud is a great way to hedge your bets and avoid getting bogged down in the mud and rain. No matter who wins the Battle of the Clouds.