In SE Asia the public cloud concept started to catch up during the last 4 years. I do clearly remember the time of meeting organizations with all levels in the organization that Immediately discard certain applications to run in the cloud due to security and compliance, serious data durability, and strict latency requirements were often considered deal breakers. It is still happening today.

Now, there are fewer stories of what doesn’t belong in the cloud

And enterprises start to ask for cloud benefits and business outcomes. Yet, only about 20% of applications now run in public and private clouds or edge environments, according to ITProtoday.

Organizations of all sizes have a choice to make about where and how to deploy apps, with hybrid cloud being a viable option in certain situations. Over the next five years, that number will rise to 80%. Given the cloud’s ability to scale dynamically, workloads that are highly variable or unpredictable or where self-provisioning is important are obvious cloud use cases.

Deciding if an app belongs in the cloud requires a careful analysis of the application and factors the options for hybrid and public clouds, relying on APIs to share data among them. Most importantly, traditional databases are often too slow and inflexible to keep up with web-scale demands in the cloud. Your cloud applications need a fast, highly scalable, highly available database platform to deliver real-time application performance no matter the amount of data or the number of users.

A modern, cloud-native data layer must be able to scale with your application, and your organization, without causing your costs to soar as well.

High availability of the data layer is also non-negotiable. So is responsiveness.

To satisfy users today, an app has to respond in about 100 milliseconds, including sending the data back and forth, which doesn’t leave much time for the database to do its thing. Finally, in a global economy, the data layer must support applications and customers no matter where they’re located, with no data loss, degradation of availability, or performance issues.

such as latency requirements, its value to the business, and more.

There are still plenty of examples of apps that shouldn’t be in the cloud.

Latency can be a critical determinant of app placement. Data processing at the edge, often based on IoT inputs, is not yet commonplace, but interest is rising in industries as varied as manufacturing, healthcare, financial services, and communications.

In those specialized circumstances, once the data is analyzed it is shuttled to the cloud for further analysis in data lakes or cloud-based data warehouses.