TPU Inference Servers for Efficient Edge Data Centers - DOWNLOAD

CNCF outlines cloud vs edge-native application principles

CNCF outlines cloud vs edge-native application principles

The Cloud Native Computing Foundation (CNCF) recently released a new whitepaper outlining edge native application principles. Having a broader swath of developers understand a common set of characteristics will theoretically help drive edge-native application development and adoption.

According to the white paper, cloud-native and edge native share several key principles, including apps and services portability, observability, manageability, and agnostic language and framework support. Applications and services are designed to abstract their coupling from underlying infrastructure to allow portability across platforms. Observability is enabled through well-documented interfaces and tool options to monitor the system and collect metrics.

Manageability is provided through interfaces and tooling options to manage resources and apps at scale, using a plugin mechanism to provide network connectivity, services and management features. Finally, developers can use a variety of popular frameworks and languages to implement apps or services.

That said, while the white paper mentions that the broad missions of cloud and edge native apps are similar, developers should be made aware of the differences.

For example, in cloud-native applications, the app model is mainly microservice components built stateless for load-balanced horizontal scaling. In edge native applications, while a service provider’s edge app may be similar, user edge apps may be singleton “monolithic”; in both cases, though, the state may be colocated with the app.

Concerning data models, centralized models backing stateless components are quite common in cloud-native applications. In edge native applications, however, caching, streaming and distributed models are often used.

The white paper says that in cloud-native apps, resilience is outsourced to cloud providers by using redundant nodes spread across failure domains. Edge native applications often rely on hardened infrastructure, with recovery architectures for stateful components.

Concerning networking, the white paper says that cloud-native apps can rely on high-speed networks with rich capabilities. In contrast, edge native apps need to account for various capacities and speeds (ranging from intermittent to poor to excellent)

The white paper also delves into specific edge native applications characteristics that include being hardware aware, external device connectivity and awareness of variable connectivity.  In terms of management, edge native applications are centrally observable while leveraging infrastructure and platform management at scale and application management at scale.

It is worth noting that the CNCF is a vendor-neutral industry organization. It is distinct from the Linux Foundation’s LF Edge effort, which is aimed at promoting the development of open source software for use in edge computing.  

Article Topics

 |   |   |   |   |   | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Featured Edge Computing Company

Edge Ecosystem Videos

Latest News