How to calculate the real cost of cloud migration

Does it make financial sense to move your workloads to the cloud? Will it be cheaper than working on a mainframe? What’s the bottom line?

Estimating the realistic cost of cloud migration and its implications, regardless of organizational size, is easier said than done.

What drives cloud migration costs?

Workloads 

The more workloads you transfer to the cloud and the more complex they are, the more expensive they are.

Data volumes 

Data transfer costs can be high in a hybrid model or with unoptimized integration patterns. One cost-efficient fix is to cluster workloads into waves, reducing data withdrawal and transit costs.

Cloud readiness 

If your workloads aren’t suitable for migration, you’ll need to tweak, upgrade or replace the applications. Here are seven migration strategies based on cloud readiness:

• Retire (decommission): An application no longer fits the future architecture.
• Retain (do-nothing): Leave the application running on-premises.
• Relocate (VM lift and shift): Stop virtualized workloads and start the VM images from on-prem virtualization environments to the cloud.
• Re-host (lift and shift — considered the most straightforward and cost-efficient strategy): Move the applications without any substantial code modifications.
• Repurchase (drop and shop): Retire an on-prem application and adopt a cloud-native SaaS alternative.
• Re-platform (lift and reshape): Lift and shift with minor tweaks to utilize cloud-native services such as managed databases in the cloud (AWS RDS, Azure SQL or Google Cloud Database).
• Re-factor (re-architect): Rework the architecture to make it more cloud-friendly.

Cloud service model 

Migrating to infrastructure as a service (IaaS) won’t cost the same as moving to platform as a service (PaaS) or software as a service (SaaS): 

• IaaS: Cloud vendor subscription fee for cloud computing, network and storage 
• PaaS: All hardware and software resources to develop, run and manage applications 
• SaaS: Vendor subscription fees with no direct cloud infrastructure fees

Cloud service providers 

Cloud infrastructure and migration costs depend on the provider’s fee structure. Typically: 

• Cloud storage fees increase the more data you store (charged /GB)
• Request fees (PUT, COPY, POST, LIST, GET, SELECT and others) are priced per 1,000 requests 
• Data transfer fees (transferring out to the internet) might include a free tier and reduce with the more data you transfer (charged /GB of data)

Most service providers provide calculators to help estimate cloud costs. Typically, public cloud is the most affordable, private cloud is the most expensive and hybrid cloud sits somewhere in the middle. Some vendors charge per-second (60-second minimum), others default to per-minute billing (specific instances support per-second billing). 

Security

Specific security controls and compliance certifications may attract additional vendor fees; for example, moving to AWS might require more advanced encryption (DSSE-KMS).

Downtime 

Cloud migration can disrupt business continuity if mission-critical systems are taken offline, potentially causing significant downtime. This is unavoidable in some instances.

Talent 

Cloud doesn’t always work like on-prem infrastructure; therefore, you’ll need to bring in fresh skills. New talent might include cloud architects and migration specialists (and possibly a development team or external partner) to prepare workloads for the cloud, execute migration and test the new environment. 

Unforeseen indirect costs can break budgets

Direct costs

• On-premises. On-prem infrastructure maintenance costs are straightforward:
  • Equipment (servers, racks, network equipment, etc.) 
  • Hardware and data-center rent 
  • Energy consumption 
  • Software licensing for operating systems and applications 
  • Labor costs 
• Cloud:
  • Provider fees for data transfer, storage, network usage and data processing 
  • Increased network connectivity (higher bandwidth)
  • Upgraded software licensing 
  • Security and compliance
  • Labor (workload preparation, migration execution and testing)

Indirect costs

• On-premises. Indirect costs are harder to calculate. For instance, a lack of modernization investment can lead to: 
  • System outages, malfunctioning hardware and loss of productivity
  • Unexpected downtime and security incidents, causing reputational erosion
  • Slow, cumbersome digital experiences that create customer dissatisfaction
• Cloud:
  • Planned downtime and having to run parallel environments
  • Staff training
  • Agile DevOps adoption
  • Data center sunk costs

Eight ways to cut cloud migration costs

Here are eight actions to help maximize cost savings and keep expenditure in check:

1.     Assess your data 

Reduce the size of suitable assets and purge unnecessary/duplicate data.

2.     Prepare your workloads 

Consolidate workloads. Processing multiple workloads in a single virtual machine (VM) is more cost-effective than running a separate VM for each workload. 

3.     Claim discounts 

Cloud vendors offer many usage discounts, such as: 

• Committed use: Use minimum cloud resources over a specific period
• Resource-based committed use: Use a specified minimum level of resources
• Spend-based committed use: Spend a set minimum in a billing cycle
• Sustained use: Use resources for a defined part of the billing cycle (e.g., more than 25% of the month) or more

4.     Check out the landing zone design 

Workloads are migrated to a landing zone. Without prior preparation, you could incur additional security and compliance costs. The setup involves aligning: 

• Account structures
• Virtual private cloud networking
• Role-based access control
• Monitoring, security and configuration management

5.     Minimize licensing costs 

According to AWS, software licensing costs can be 300% more than the total cost of running cloud workloads. Open-source alternatives to proprietary solutions (especially in database and server management) can help minimize such costs. 

How to calculate cloud migration costs

Cloud migration costs fall into two bands: 

• Cloud: Data transfer costs plus future storage, network and processing
• Human: Qualified internal/external labor

Alternatively, you can break costs down into three phases: 

• Pre-migration: CapEx elements to be eliminated. Workload re-platforming and re-factoring where necessary. Data restructuring. Network and security redesign
• Migration: Cloud consumption. Migrate, integrate and test workload resources. Migration software licenses and consulting services management
• Post-migration: Cloud-vendor fees for storage, network usage and data processing. Auxiliary services — hardware security modules for encryption, DDoS protection, observation and monitoring. Labor (hiring, training). Software licensing — security information and event management (SIEM) tools

Not forgetting the sunk expenses for decommissioned on-prem infrastructure. 

6.     Use cheaper regions

Server location can affect vendor fees, so maybe consider a single-region setup in a less expensive area.

7.     Consider mainframe hybrid cloud

A mainframe hybrid cloud model can help you benefit from public and private clouds (flexible and cost-efficient) while retaining your mainframe’s superior security and compliance capabilities.

8.     Monitor resource utilization

Post-migration, adjust for consumption to avoid overspending. Identify and neutralize idle resources, such as unused persistent disks, IP addresses, custom disk images and VM instances. Decommission environments and resources no longer needed. 

Connecting the dots

The cost control required to maintain expected performance and minimize disruption requires specific cloud migration expertise. So, first things first, analyze your in-house talent to determine whether you need an external partner, such as DXC.

Our domain experts conduct a comprehensive inventory and analysis of all workloads and dependencies to evaluate the project, prevent scope creep and maintain budgetary discipline.

We ensure that regular cost reviews are carried out and spending is optimized through utilization monitoring, instance rightsizing, eliminating unused resources, managing software licenses and organizational changes. Our teams also account for the sunk costs associated with excess labor, vacated data centers and idle hardware/software. Nothing is overlooked.