Updated on 2025-05-07 GMT+08:00

How Do I Choose a Cutover Solution (With or Without Downtime)?

Service cutover is a crucial step in the entire cloud migration process. Unexpected situations that occur during a cutover, if not handled promptly, can directly impact business continuity. Different services have different requirements on service continuity. Some services cannot tolerate even minimal downtime, as any interruption leads to significant losses. Some can be stopped during cutover. For example, OA systems may be stopped and upgraded during non-working nighttime hours without major disruptions. For some others, only browsing services must remain available during cutover to ensure a smooth user experience, while write operations can be temporarily halted. When designing a cutover solution, it is essential to assess workloads and their tolerance for downtime.

Three cutover solutions are available: cutover with downtime, cutover with writes stopped but reads continued, and zero-downtime cutover. The table below compares different cutover solutions.

Table 1 Comparison of cutover solutions

Cutover Solution

Description

Data Consistency Risk

Service Change Cost

Downtime (Hours)

Read

Write

Cutover with downtime

Commonly used solution that can ensure data consistency

Low

Low

0.5~3.5

Cutover with writes stopped but reads continued

Less used solution, which requires service changes to implement cutover with writes stopped but reads continued. This solution can ensure data consistency.

Low

Medium

Continued

0.5~3.5

Zero-downtime cutover

Rarely used solution, which requires service changes to implement dual-write and bidirectional data synchronization. This solution requires service changes to ensure data consistency, which can be fairly complex.

High

High

Continued

Continued

Each solution has its advantages and disadvantages. There is no perfect option that offers low risk, low cost, and minimal downtime. Businesses need to select their cutover solution based on service scenarios, tolerance of downtime, and cost-effectiveness. Consider the following when choosing your cutover solution:

  1. By industry

Different industries have their own standards and requirements. For example, retail e-commerce platforms typically have significantly fewer transactions or even zero transactions during early morning hours. For them, performing a cutover with services completely stopped during early-morning hours is acceptable. However, for some other industries, such as ride hailing apps that must be available 24/7, stopping services is simply not an option. For them, a zero-downtime cutover is the only viable option.

  1. By service importance

Mission-critical services, such as online games and financial services, must be available 24/7. For them, a zero-downtime cutover is the only viable option. Other less critical service, such as OA and operations systems, can tolerate brief interruptions. For them, you may choose cutover with downtime.

  1. By project period

If a full-service stoppage is not an option, consider using a zero-downtime cutover solution, if there is sufficient time. In this case, some service changes may be required. For example, an active-active design may be needed. Conversely, if the cloud migration project is on a tight schedule and the customer cannot devote enough manpower to the project, choose cutover with downtime, as it requires minimal changes to the migrated service systems and minimal manpower from the customer side.

  1. By cost-benefit ratio

A zero-downtime cutover often requires significant R&D efforts to adapt service systems, while a cutover with downtime minimizes such investment. Businesses must carefully weigh the costs against the expected benefits.