Businesses use cloud computing. In reality, most enterprise organisations (as well as small to medium-sized businesses) tapping into the flexibility and control benefits of service-based computing and its ancillary data functions use more than one type, strain or instance of cloud computing.
Of course, we’re talking about multi-cloud. This is the implementation of services from more than one Cloud Services Provider (CSP), which may characteristically also mean a selection of different cloud optimisations from each CSP.
But this is not a ‘how are clouds different?’ story, that tale has already been widely told and you can read all about public, private, hybrid, multi, poly and armadillo cloud if you want. Spoiler alert: there is no armadillo cloud, please don’t look for it, at least not yet.
So let’s move on. What we need to know is what kind of yardstick the market uses to tell us how far the multi-cloud revolution is happening and how should we approach multi-cloud deployment in the first place?
A multi-cloud yardstick
Cloud SaaS specialist Wanclouds is typical of many firms in pointing out the general multi-cloud trends that appear to be happening. The company is an implementation partner with AWS, Google Cloud and IBM Cloud and has used its experience at the coalface to produce its 2022 Cloud-Native Trends Outlook by surveying 416 US and UK IT decision-makers in Q4 2021.
Wanclouds found that 84% of IT decision-makers say they use more than one cloud platform today. Most commonly, they said they use two (29%) or three (19%) cloud platforms.
There are obviously plenty of benefits of the multi-cloud approach. Most notably, companies can focus on using the best cloud environment for the specific needs of each workload. But perhaps the most cited reason by IT executives is avoiding vendor-lock-in with any cloud provider.
The multi-cloud pain point
But to say that two-thirds, three-quarters, or in this case 84% of firms use multi-cloud is hardly progressive or revolutionary. We know this happens. What might be more insightful is to know what the knock-on effect of ‘needing’ to be multi-cloud is.
CEO of Wanclouds Faiz Khan explains that many of the customers he talks to are building multi and hybrid cloud environments, but they’re increasingly running into longer project runways and the need for bigger budgets. Most firms typically cite around one to two months for a single multi-cloud application migration; but in multi-cloud, that same migration takes up to six months. In these latter more complex scenarios, spiralling post-migration cloud costs and billing system complexity confusion are rife.
“The pain of getting hit with unexpected cloud costs is something our team regularly hears from organisations of all sizes and across every industry,” said Khan.
“IT teams have found themselves helpless against hidden charges and a lack of company-wide cloud visibility within their platforms’ billing systems. This new data showcases how widespread these issues are and suggests that cloud providers still have a long way to go to deliver a user experience that meets the evolving needs of their customers,” added the Wanclouds CEO.
Cloud migration transformation
At the risk of using too many tech industry buzz-terms in one mouthful, it seems clear that cloud migration needs transformation.
Clearly, making the decision to implement a multi-cloud strategy is difficult. From the decision to pursue a multi-cloud journey to defining the requirements to laying the foundation to identifying and deploying applications and services to the multi-cloud environment, the process requires a solid strategy and flawless execution to succeed.
This is the opinion of Hamilton Yu, CEO of Taos, a cloud migration and application modernisation company that now lives under the ‘an IBM company’ family umbrella.
"Despite an uptick in adoption, multi-cloud strategies are thought of as a nice to have an option rather than a transformative business option,” said Yu. “The benefits of such a deployment far outweigh any concerns an organisation may fear. However, it is important to properly plan and carefully consider options to get the most out of a multi-cloud environment.”
The Taos advisory team say that a true multi-cloud environment is designed to have applications and services deployed on disparate resources with an existing network to route between locations.
As we have already suggested, this deployment format allows organisations to use the strengths of different CSPs to meet their needs for any particular data set or application. It also enables firms to choose between cloud locations for better performance, to satisfy geographically-based residency requirements; or ensure business continuity.
In actuality, most of all, multi-cloud often just helps to avoid vendor lock-in, so how we do it?
7 steps to multi-cloud
Yu and team offer seven steps to multi-cloud migration that can help businesses in any industry vertical to move forwards. Briefly summarised these are:
- Rationalise requirements - Gather both business and technology requirements and develop a decision tree, taking into account which CSP can solve which requirements.
- Create connectivity - Prior to the first application or service deployment, develop a global network strategy to manage connectivity between clouds.
- Viable visibility - Design solutions that work between clouds to bring security visibility with scanning tools to identify misconfigured cloud resources and services.
- Good governance - Build governance to introduce consistency and order into an otherwise disorganised environment.
- Thorough training - Not every administrator will know every cloud, offer training to manage the entire multi-cloud environment across skillset limitations.
- All-in automation - When automating multi-cloud deployments insist upon tools that work across all CSPs, such as Terraform offer a consistent mechanism for deploying infrastructure and provisioning services.
- Talk timing - Work closely with the finance team to analyse the total cost of implementing a multi-cloud strategy and work out what happens first.
Rushing to implement a multi-cloud strategy will result in inefficiencies, rework and in some cases, a total replacement of the work performed.
The lessons here are straightforward enough, yet as of 2022 we would carefully and respectfully like to suggest that they still need saying out loud. The 7-point list above is probably not an exhaustive safely checklist that will apply to every in-flight multi-cloud journey once it takes off, but it’s an arguably pretty good set of pillars upon which to start the journey.
