How To Build Applications The Edge Native Way

I’ve been involved in some great edge-related activities over the past few years. Some of these activities have been focused on Infrastructure as a Service (IaaS) or orchestrating workloads around a distributed cloud. While IaaS and workload orchestration are important parts of an edge cloud solution, where does the platform for developers fit into all of this?

In my last blog post, I said that building applications for the edge will require a different set of tools. The edge (and the telco edge in the most extreme sense) is about real-time performance and an improved experience for humans and machines. It’s this concept of service immediacy that goes beyond industries.

Even though there are many applications and technologies residing in the centralized clouds, they don’t always translate into workloads that make sense at the edge. The main reason a traditional cloud native approach doesn’t work is because it doesn’t face the same barriers as distributed architectures. Even if they technically can run at the edge, that doesn’t mean that they’re optimized for the edge. 

5G, Telcos, and building for the edge

Telcos, especially mobile network operators with 5G, are transforming their business in two main ways. The first way is technical, i.e. network transformation and evolving from the traditional physical network functions approach to network deployment to a virtualized approach (VNFs and now CNFs). The second way is monetization, i.e. business transformation and how they monetize that network asset and capabilities to attain a return on their investment.

Telcos are essentially becoming edge technology services with the goal being to offer their unique value proposition in the form of network-centric edge cloud services.

Telcos that focus their edge cloud services on real-time performance and service immediacy are going to be more successful and fit into the technical and business requirements of modern distributed systems.

What application developers need for the edge

Rather than simply hosting application workloads created by others to run at the edge, 5G mobile network operators can become enablers of solutions. But first, they’ll need to address the needs of enterprise application developers at the edge. Telco edge cloud providers need to consider three main concerns: Performance, developer velocity, and costs. 

Performance is all about dynamic scalability, low latency real-time insight, and actuation. Developer velocity isn’t just about the speed of development or deployment, it’s about the rate of developer adoption as well. The costs also have to be worth the investment – does the platform’s cost result in tangible performance and increased developer velocity?

To meet the demands of enterprise applications and the developers that create them, network operators need to support a polyglot approach to enterprise architecture primitives. Imagine a platform that embraces an edge native approach to application development. Developers can leverage stateful serverless computing to maximize the dynamic, mobile (migratory) nature of deployed software services at the network edge.

How Macrometa overcomes obstacles of the edge

The edge is born from the principle that data has gravity. Basically, the existence and mass the data has, in itself, attracts processes and workloads to it. Like any large body that has gravity, other objects are attracted to it and sit (and therefore operate) in its orbit.

At Macrometa, we have built an edge cloud platform that embraces the principle that data has gravity. For example, if insights need to be generated in real-time, our platform enables a distributed approach to application design that brings workloads, AI/Machine Learning, and application logic to where the data actually is. The mechanisms required to intercept, capture, process, and manipulate data have everything they need.

Telcos and mobile network operators that work with Macromenta are able to deploy an application that is natively low latency, built for high-velocity data ingestion, manipulation, and actuation. Data replication, query, processing, and insights can be performed and generated across the operator’s footprint with near-instantaneous synchronization, or even globally (off-net) within ms.

Macrometa’s streaming engine is built around persistent streams and real-time stream processing via our Stream Worker technology. Our Stream Workers allow developers and Telco enterprise customers to define sources and sinks, tables, processors, and javascript functions. Stream Workers are extremely powerful in real-time event execution in an ETL / ELT architecture. Stream workers give the developer the power to collect, shape, analyze and act on application in-flight data and data at rest.

The platform also provides a stateful container runtime that is distributed across a network topology with both compute and FaaS capability.

Hopefully I’ve managed to explain what we have in our platform, its features, and how our Telco customers are using it to offer real-time edge cloud services. It’s pretty exciting stuff.

We (Macrometa) have telco customers today who are right in the middle of this transformation and using the Macrometa platform to build their edge cloud services while addressing complex use cases like connected vehicles, gaming, and smart cities, among others.

If you head over to Macrometa Docs you can find tons of demos and tutorials, from simple “hello-world” type applications to more complex ones like Crypto Trading, and Fleet Management. We also have all the documentation you need to get started to familiarize yourself with our platform and build your own global edge applications today.

Request a 30 day trial.

‍

Photo by Uriel SC on Unsplash

Further reading

Edge computing vs cloud computing

IoT Infrastructure Guide

Introducing GeoFabrics