Skip Rehost: Replatform First to Cut Costs and Avoid Cloud Waste

Stop Rehosting Everything: Replatform First

You've heard the advice: lift and shift to the cloud, and you'll save money fast. But that's a trap. Rehosting is the fastest way to move, but it's also the fastest way to burn cash on cloud waste. The smarter play is to replatform first—make small, targeted optimizations during migration—and save your rehosting for the few apps that truly need it. Here's how to do it, step by step, based on real-world case studies and cloud provider guidance.

Who This Is For

This guide is for you if you're planning a cloud migration and want to avoid the classic mistake of moving everything as-is, only to discover you're paying for idle resources. You're probably under pressure to show quick wins, but you're willing to invest a little more upfront to reap bigger savings later. If that's you, read on.

Step 1: Assess Your Portfolio – Find the Zombies First

Before you touch a single server, do a thorough assessment. Azure's migration framework starts with an Assess stage, where you create a full inventory and dependency map of your servers, services, and apps, and estimate cost savings using a TCO calculator (Azure Migrate). This isn't busywork—it's how you avoid missing critical dependencies that will derail your migration later (DigitalOcean).

But here's the kicker: while you're inventorying, look for zombie applications. AWS defines zombies as those with average CPU and memory usage below 5% over 90 days, and idle apps as 5–20% usage with no inbound connections. These are prime candidates for the Retire strategy (AWS Prescriptive Guidance). In a case study, AWS helped a customer consolidate 80 SAP systems and achieve 30% cost savings by retiring unused or underused instances (AWS). You can do the same: if you find apps that fit that profile, don't migrate them—kill them. You'll save money before you even start.

Step 2: Decide What to Replatform – Not Everything Needs It

Once you've retired the zombies, you'll have a cleaner list. Now, resist the urge to rehost everything. Replatforming, or 'lift, tinker, and shift,' makes minor optimizations without changing core architecture—like moving SQL Server to a managed database service (DigitalOcean). AWS gives concrete examples: moving SQL Server to Amazon RDS, using Graviton processors, or porting .NET Framework to .NET Core on Linux (AWS Prescriptive Guidance). These tweaks can cut your compute costs significantly because you're not paying for licensing and management overhead.

But not every app needs replatforming. Some are fine as-is. You'll have to make judgment calls. In a Red Hat survey, replatforming was the most common migration approach at 20%, while other strategies ranged between 10% and 19% (Red Hat). So it's a popular choice, but not a one-size-fits-all.

Step 3: Plan Your Migration – Use a Phased Approach

You need a plan that balances speed and safety. AWS recommends a three-phase approach for large migrations: assess, mobilize, and migrate, which lets you create a plan, set up a foundation, and migrate the first set of apps in weeks (AWS Prescriptive Guidance). Azure uses four stages: Assess, Migrate, Optimize, and Monitor (Microsoft Learn). IBM has six steps, including selecting the provider and designing the target architecture (IBM). Pick one and stick to it.

For a case-study approach, start with a pilot workload that's a good replatform candidate. For example, take a .NET Framework app that you'll port to .NET Core on Linux. This gives you hands-on experience with minimal risk. Then, scale up.

Step 4: Watch Out for Cost Creep – It's Sneaky

Here's the warning: your cloud bill will balloon if you're not careful. Flexera's 2025 report found that 84% of organizations struggle to manage cloud spend, and cloud budgets exceeded limits by 17% (Flexera 2025). Wasted spend on IaaS and PaaS rose to 29% in 2026, the first increase after five years of decline (Flexera 2026). Rehosting makes this worse because you're moving workloads without optimizing them, so you're paying for idle capacity.

Replatforming helps because you're right-sizing as you go. But you still need to track your spend. Flexera found that 87% of organizations now use cost efficiency as their top metric for cloud goals, and 64% measure value delivered to business units (Flexera 2026). So, set up a FinOps team early—63% of organizations already have one (Flexera 2026). They'll help you avoid the trap of paying for more than you use.

Step 5: Execute Migration – Do It in Waves

When you're ready to move, don't do it all at once. AWS notes that rehosting allows large numbers of machines to be migrated without long cutover windows (AWS Prescriptive Guidance), but replatforming takes more time per app. Use a wave approach: group apps by dependency and move them in batches. For each wave, you'll assess, replatform, test, and cut over.

For example, if you have a web app that relies on a SQL Server database, replatform the database to a managed service first, then move the app. This way, you reduce risk and can measure the impact of each change.

Step 6: Optimize and Monitor – The Work Never Ends

After migration, you're not done. You need to optimize continuously. Azure's framework includes an Optimize and Monitor stage (Microsoft Learn). This means right-sizing instances, turning off idle resources, and using serverless where it makes sense. Red Hat notes that serverless is an event-driven model where an idle function costs nothing (Red Hat). That's perfect for spiky workloads.

Also, keep an eye on your security posture. Remember, under the shared responsibility model, you're always responsible for your data, endpoints, and access management (Microsoft Learn). Don't let security lapses negate your cost savings.

What Can Go Wrong

Here's the biggest pitfall: you skip the assessment and replatform everything, only to find you've replicated your on-premises issues in the cloud. Or worse, you rehost everything and end up with a massive bill. In one case, a customer exited 17 data centers in 30 months using AI (AWS), but that's not typical. Most organizations struggle with data security, compatibility, and under-assessing the portfolio, leading to missed dependencies (DigitalOcean). To avoid this, invest in dependency mapping and pilot testing.

The Single Most Important Thing to Remember

Replatform first, rehost only when necessary. It's the sweet spot between speed and cost efficiency, and it sets you up for long-term cloud success. Don't let the allure of a quick lift-and-shift blind you to the waste you'll leave behind.

Sources

• DigitalOcean - https://www.digitalocean.com/resources/articles/cloud-migration-strategy
• Red Hat - https://www.redhat.com/en/blog/how-should-you-modernize-your-applications
• Microsoft Learn - https://learn.microsoft.com/en-us/training/modules/design-migrations/3-describe-azure-migration-framework
• AWS Prescriptive Guidance - https://docs.aws.amazon.com/prescriptive-guidance/latest/large-migration-guide/migration-strategies.html
• Flexera 2025 - https://www.flexera.com/about-us/press-center/new-flexera-report-finds-84-percent-of-organizations-struggle-to-manage-cloud-spend
• Flexera 2026 - https://www.flexera.com/blog/finops/the-new-era-of-cloud-what-2026-data-tells-us-about-spend-scale-and-strategy/