The terms “hybrid” and “multi-cloud” are tossed around very loosely within our industry, I’ve been guilty of this myself. But there’s no denying hybrid and multi-cloud networks are taking hold rapidly in enterprises of all sizes. And with these networks come different and competing schools of thought on how they should be built, scaled and managed. Should IT organizations follow a vertical approach by building multiple clouds to support individual applications? Or should they consider building their networks horizontally, spreading the applications and the management of those applications across multiple clouds, resulting in best-in-breed cloud services and consistent policies and protocols? If you read the title of this article, you know which side I come down on.
Let’s start by clarifying what it is we’re talking about when we talk about hybrid and multi-cloud networks. Are we referring to hybrid multi-cloud at the application layer? Or are we referring to hybrid multi-cloud in the context of running multiple clouds in support of individual applications? Without making that distinction, it's difficult to adequately address the need for companies to adopt a horizontal approach.
Many IT organizations currently pursuing their hybrid cloud or multi-cloud strategy take a vertical approach by building multiple clouds to support individual applications. For example, a company may consume cloud services from Amazon, Google, Microsoft and others. In each of these instances, the application stack is entirely contained within the cloud itself. While those organizations and the application teams may be well versed in those cloud technologies, when you take a step back you see each team is operating in separate models and functionalities – having to build complete technology stacks for each application running on each cloud – resulting in gaps across teams, unnecessary costs and inefficient business practices.
Horizontal benefits outweigh challenges
In a true horizontal hybrid or multi-cloud approach, splitting the application stacks across multiple clouds gives organizations the ability to uniformly deploy infrastructure in support of all the applications they would otherwise have to configure bespoke solutions for in a multi-cloud environment. For example, an organization can have its database on Oracle Cloud, put its CDN on Cloudflare, move its web front end to Amazon, and get DNS services from Google. This is a true hybrid or multi-cloud implementation. And when it’s time to scale your network one way or another, to implement new policies and services, or to audit your network, it becomes very easy because you only have to do it once versus coming up with per application stack configurations.
For all its operational benefits, a horizontal approach isn’t without its challenges.
A large reason horizontal hybrid and multi-cloud infrastructures have struggled to gain traction throughout the enterprise is that application teams are still very tied to their cloud technologies. They’re familiar with them and they know how they work. To them, it makes sense to stay with the tried and true.
Still…when executives get the bill at the end of the month and question why they're paying a premium for certain services. “What about diversification?” they ask. “Make the clouds compete and keep each other in check.”
This creates a tug of war where executives are looking to keep costs down, the app development teams don’t want to let go of their technology, and the network teams are caught in the middle trying to make sense of everything. They’re already being challenged by the complexities of today’s infrastructure and trying to accomplish everything they’re being asked to do without finding some scaling properties in the design itself – scaling properties that a horizontal configuration can provide.
And of course, there is the ever-present growth of 5G technology. As the industry begins to leverage 5G, we’re seeing a number of organizations at least entertain the idea of divesting from centralized cloud architectures and leveraging 5G to push applications far out to the Edge. This is going to create tremendous pressure on infrastructure teams to maintain efficient, optimized running infrastructure in support of key, business-critical applications.
Adoption requires aligned technology and business approaches
Without question, the adoption of hybrid and multi-cloud networks is becoming more of the norm for organizations. But different segments and industries are moving at different paces, with cloud-first or cloud-native companies moving at warp speed. More established industries are struggling with this digital transformation, primarily because they still have so many legacy, snowflake configurations that simply makes it more challenging to move towards hybrid and multi-cloud.
For organizations to successfully adopt a horizontal approach, two things need to happen.
First, it has to be a top-down decision: a strategic technical approach aligned with the business strategy. Second, organizations must tear down silos and bring application development and infrastructure teams together. That’s when they can see the true benefit of hybrid multi-cloud and the pattern starts to shift.
Looking towards 2022, IT leaders will be focused on successfully aligning their hybrid and multi-cloud initiatives for increased operational efficiency. In doing so, more should lean towards a horizontal approach, uprooting traditional processes across all departments, and deploying cloud-native tools on a wider organizational scale to reduce time and stress.
More in Cloud & Hyperscale
Conference Session Panel: How will cloud and colocation need to evolve in an Indian context?