Until around five years ago, if someone had asked me about multi-cloud, I would have responded with "eh .. its just a hype with no real value, a waste of money and time. No sane person should even consider going there." Primary reason behind this thinking was:
- It is tough to operate and get the skillset.
- This approach forces an organization to settle with the "least common denominator".
Fast forward 2021, while working with several organizations within US and Australia, I figured that multi-cloud is becoming a reality. I thought I would have a "multi-cloud" blog series where I can discuss why organizations end up adopting multi-cloud, and perhaps compare in details some of the providers' findings when it comes to their individual offerings, how they differ, challenges etc.
MultiCloud by accident
I have seen businesses with a very strong relationships with Microsoft. They have Enterprise Agreement, license portability options, getting hefty discounts when migrating to Azure. Even then, some of their critical workloads land in Amazon. Why? Its shadow IT! Business lines want their product in front of the customers quickly. The developers are approached and they find AWS to be very dev friendly. They get the company credit card and create account(s) and start putting the pieces together without a any reference architecture in place, without discussing their approach with architecture - and Voila! product ready for a launch. Business is happy! What happened to the tech strategy? Ignored. Can this situation be avoided? Yes, if in this example scenario, the foundational element such as automated governance of Azure - policies, account/subscription management, cost management, authentication and authorization - was ready to go. It would not have given developer a chance to think about any other platform but could easily "consume" the Azure platform services.
On numerous occasions, I encountered companies subscribing to variety of SaaS and PaaS solutions like Salesforce, Power BI, Office 365, Snowflake, etc. They already establishes this multi-cloud environment unknowingly but sadly without a proper strategy/foundation in place.
MultiCloud by strategy
A well defined strategy could lead to a multi-cloud adoption where an organization can decide to adopt this strategy based on:
- Value addition: A cloud provider has specific strength, their key value proposition. Depending on the organization culture, strength, technical/operational ability and processes, workload, one can choose to run a specific workload on a specific provider.
- Backup: An organization can choose to adopt one specific cloud just for backing up and securing enterprise data.
- Dev/Test vs Production: Experimenting with one cloud provider while running productive environment in another.
- High Availability and DR: Expensive and comes with an overhead of management and network strategies, this approach also comes with a great skillset requirements to operate.
Whatever might be the reasons, organizations running multi-cloud would do well if they can create a strategy to operate the environment securely and cost-effectively. There is a multitude of tools and technologies to choose from: starting from network, security, deployment, compute and databases, serverless, etc which are cloud-agnostic. But most importantly, the org processes and culture have to be "multi-cloud-ready".
My next blog will discuss variety of strength and shortcomings from the "big-3".
Top comments (0)