Cloud-Native Architecture

A cloud-native architecture is one in which an application has been designed specifically for cloud deployment (contrast to “lift-and-shift”, where applications are moved to the cloud unchanged.)

Cloud-native architectures are those that take advantage of the things that cloud platforms do well. For example, it is easy to scale out a cloud-based application by simply adding more server nodes, as well as to scale back down if the scale-up was due to a temporary spike in demand (for example, a retailer’s website on Black Friday or Cyber Monday). So cloud-native applications should provide horizontal scaling capability, and not be monolithic apps that are deployed on a single server. This is done in part by adopting a microservices architecture, where the app is decomposed into small functional pieces that can be deployed independently and connected via stable, clearly defined APIs.   

Similarly, once an application is scaled out and deployed across multiple nodes, it is of course desirable that the system be resilient to failures affecting one or a small number of nodes. So a cloud-native application should be designed to continue to perform in the event of a node outage and to recover gracefully from failures. 

Developing capabilities such as horizontal scalability and fault tolerance is tough to do from scratch, but these features are a key component of containerization platforms such as Docker; this is why many companies will move to a containerized platform as part of any migration to the cloud, and why the adoption of a container-based architecture is considered synonymous with being cloud-native.  A container orchestration framework like Kubernetes (which is the basis of packaged stacks like Red Hat OpenShift) to manage the deployment of containers.

Cloud-Native Architecture Diagram

A container-based architecture takes the application that is to be deployed and bundles it together with all of its dependencies into an image that includes everything needed to run the app — language runtimes, software libraries, and the application code itself. This image is then stored in a centralized repository (public or private), and deploying the app is then done through simple tools that copy the image to the target system, create a container for it to run in (which provides isolation from other containerized apps that may run on the same physical server), and makes connections as needed to resources such as network ports.

While there isn’t a universally agreed-upon set of criteria to define a cloud-native application, most sources agree that an application that has been decomposed into a set of microservices, with these microservices deployed via containers, qualifies as cloud-native. Having a continuous integration/continuous deployment framework that allows the container images to be updated and deployed regularly in accordance with agile software development practices is another characteristic that is commonly associated with cloud-native architecture.

Related Topics


Microservices Architecture


Relevant Resources

| Video
| 60 minutes

How to Deploy Cloud-Native Applications in Kubernetes

Are you planning a cloud-native application or looking to deploy Hazelcast in the cloud? Then this webinar is for you.

Kubernetes is a container orchestration system that has seen a meteoric surge in popularity.  It works with a range of container tools including Docker, or as a part of your PaaS platform, such as Red Hat Openshift Container Platform, Cloud Foundry with its Project Kubo, and Google’s KNative platform.

In this webinar, we will take a look at what it means for anyone currently using or interested in using Hazelcast to build a cloud-native application. We will review how to setup and deploy using a sample application, and get you ready to have a starter application of your own running on Kubernetes.

White Paper

Cloud Migration and the Role of In-Memory Technologies

Hazelcast Cloud delivers enterprise-grade Hazelcast software in the cloud, deployed as a fully managed service. Leveraging over a decade of experience and best practices, Hazelcast Cloud delivers a high-throughput, low-latency service that scales to your needs while remaining simple to deploy. If you’re considering moving to the Cloud, or are looking for an easy ramp on deploying in-memory technology, this white paper on migrating in-memory to the cloud is an informative and helpful resource.

View All Resources