How to lay foundations for cloud at scale

You also need to consider how you’ll provision and manage resources in the cloud. When early decisions are shaped by long term perspective, you’re more likely to enjoy a secure and seamless transition that quickly unlocks cloud benefits.

That said, adopting cloud at scale is a mammoth undertaking. It can be difficult to figure out where to start and how the different technical elements hang together. This blog post looks at how to bring clarity to the early planning stages, the benefits of application-centric migration and cloud deployment patterns.

Hold an exploratory workshop

We’re big fans of exploratory workshops which define goals and align expectations at the start of a major IT project. Key stakeholders and subject matter experts get together to determine what’s needed, as well as why, when and how it needs to happen. 

From a cloud migration perspective, this is an opportunity to establish whether the primary driver is cost savings, staff productivity, operational resilience or business agility. To make the most of the cloud environment, and deliver on these goals, modernization is typically essential. But if there’s a datacenter renewal approaching or a hardware refresh on the cards, speed will be a critical factor. In most cases, compromises have to be made. It’s helpful to acknowledge this at the outset, as it sets the tone for the entire process.

Exploratory workshops also create space for open conversations between cloud experts, application owners and business leaders. This is important because it allows potential issues and constraints to be highlighted upfront.

One area to focus on is the in-house capabilities and capacity for cloud deployment as well as ongoing management and operations. Many organizations assume that traditional approaches for maintenance, security and governance will work in the cloud. Much of the time they don’t. Understanding this enables well-informed risk-based decision making about how to handle different aspects of the migration. 

Whatever your goals, a well-structured exploratory workshop, ideally facilitated by an experienced cloud consultant, will ensure a strong start.

Devise an application-centric approach to migration

A few weeks ago, one of our cloud solutions architects wrote about three situations where a lift-and-shift is right, and three where it isn’t  . We won’t delve into the pros and cons of lift-and-shift versus refactoring, re-platforming or rebuilding here. However, as he indicates, it’s important to consider each application individually when deciding the best route to the cloud.

This is where the benefits of holding an exploratory workshop come to the fore. It means every application can be considered though the lens of guiding principles that have been agreed in advance. Potential challenges and opportunities can also be assessed pragmatically according to what has been discussed.

For instance, if an application involves sensitive data, it might be prioritised for refactoring in line with well-architected best practice. This will allow engineers to leverage cloud-based tools and services that enhance security and ensure compliance with relevant industry requirements. On the other hand, refactoring might take too long and present too many technical challenges. In this situation, it might be better to set a date to retire the application and replace it with a cloud-native alternative. Being able to make decisions like these with speed and confidence helps drive more efficient and effective progress.

Determine the most appropriate cloud deployment pattern 

Cloud deployment patterns underlines the need for a planned approach to deployment when adopting cloud at scale, to avoid what it terms ‘a free for all sprawl’. This can occur when users are allowed unrestricted access to services and resources in the cloud. It potentially results in issues with cost visibility and management as well as difficulties managing and evaluating security and compliance.

There are five distinct deployment patterns: service catalogue, infrastructure module pipeline, inline validation pipeline, freedom with guardrails and implementing control strategy. However, it points out that patterns can evolve or be supplemented as requirements change or team capabilities develop. Many organizations start out with a prescriptive approach which becomes more flexible as teams become more familiar with cloud security best practice and well-architected principles. It’s also possible to adopt multiple deployment patterns, so application teams with varying knowledge and skill levels can work at their own pace. 

Something we advocate is the use of landing zones which provide a consistent template for cloud deployment after migration. They offer a baseline for critical elements such as governance, data security, identity and access management, network design and logging. They’re configured with organizational policy and industry requirements in mind, which makes them a valuable part of the mix for large enterprises in highly regulated sectors.

The planning and preparation sweet spot

You need to plan for largescale cloud migration. But you also need to avoid over planning. With such a complex task, over planning is likely to hinder progress or result in analysis-paralysis. Instead, we find it’s more effective to establish goals and expectations at the outset, using them to shape high-level planning. This informs technical decisions about the migration, modernization and deployment of applications during the longer-term cloud journey. It’s about facilitating a stepwise approach where every decision is based on where you are, where you want to be, and a realistic understanding of what it will take to get there.