Intent-driven Edgeification

As you lead your organization through a digital transformation, do you worry about the next wave of complexity of moving your services to the edge? Avesha’s intelligent, edge-aware & intent-driven platform is here to help.

Avesha
4 min readFeb 5, 2021

Imagine you are driving a bus full of people on a racetrack. Suddenly, you have been called on to win a race against a fleet of Formula1 race cars. The challenge of transforming our bus into a race car, all the while staying competitive and not stopping, is daunting.

Any engineering leader who has had to migrate applications / solutions to the cloud (while simultaneously evolving from monolith to microservices) knows this challenge all too well — tools and technologies evolve faster than enterprises can adapt. Aha! never forget operating them at scale.

The transformation landscape

Dark logic — the tribal knowledge and dependency spaghetti of current monoliths — obscure so many of the relations. The dependency of data, its availability, retention capability, and interdependent messaging built over the years compound the challenges of transformation.

While undertaking this 2-dimensional transformation, consumers of the system do not care about your transformation journey; they care about features, availability, and efficiency of the system. The above landscape is real and challenging.

While we travel through this complex transformation of breaking into microservices, segregating some of the basic functions and adopting services meshes, it is imperative upon us to not forget “Edgeification”. Edgeification is the term we use to denote distribution of workloads from cloud to edge. The industry can’t handle yet another wave of transformation without some respite from the challenges they already face. Enterprises need to focus on customer value; NPV (Net Present Value) of an enterprise is not calculated based on how many transformations we have gone through, but what value we have delivered to the customers and their adoption of our services delivered.

“Third Act of the Internet — State of the Edge 2020”

In the ‘2020 State of the Edge report’, described as the ‘The Third Act of the Internet’, edge is real. Today’s SaaS providers have global reach of customers. Customer behaviors produce lot of meaningful data closer to customer presence. Hauling data to a centralized cloud where we brought our application to utilize elastic compute will not serve customer experience. Bringing compute and application to where data is generated, will help drive customer experience. Disaggregating or distributing application presence closer to customer needs would require a mesh framework which handle application intents, starting from application connectivity, data governance, security and traditional infrastructure centric services. At Avesha we believe “Edge” is defined where data should not be long-hauled and maintain customer experience.

Source: State of the Edge 2020: A Market and Ecosystem Report for Edge Computing

Builder Experience deploying/building Disaggregated Application

While an application builder embraces DevOps culture and practices extreme automation (IaC, Shift-left testing, policy driven scaling) many deployments per day is a common practice in many enterprises. Organizations which have adopted containerization, declarative configuration and automation have benefited from consuming the latest technologies. At Avesha we are helping in extending that acceleration by introducing ‘Application Intent As A Code’ in improving velocity.

Application Intents

What is an intent? Per English dictionary and Adjective: ‘Determined to do’.

Application builders have great tools like Containers, Kubernetes, Service Mesh which help automate a lot of application compilation dependencies, configuration, deployment, and self-heal infrastructure centric intents. These are a fleet of services that bring enormous momentum to change your applications to a microservices based architecture.

Distributing such application clusters across edges and connecting them add networking challenges coupled with application data movement challenges.

An application embodies many ‘Intents’ for completeness of its business logic function. There are many ‘Application Intents’ which are common across all services aka microservices. Like ‘CRUD’ (Create, Read, Update, Delete) are basic functions of storage or backend applications, assign compute where actionable data exists, synchronize data for analytics at a central place where data lakes exit, distribute the workload where user engagement is better served: these are ‘Application Intents’ which need to be performed for many microservices with intelligence based on the situation, location, and end-user behavior. In order to achieve such functionality and be focused on delivering microservices that are purely focused on business logic and to limit the number of microservices dependencies, would there be a desire to have a platform that supports a framework of descriptive language like YAML expressing Application Intents’?

Think about it Does this not unshackle you from repeating or creating a dependency on microservices? Does this not improve your velocity of delivering the services faster? Does it also help you in proper sizing a microservice? Does this not offer comfort in building your microservices if the underlying ‘Application intents’ are hardened and built on zero trust principles? Does it help enhancing customer value if such microservice placement is intelligent and is edge-aware? Does it also help if you can describe one of your ‘Application Intents’ to include latency expectations for your microservice and such policies are marshaled and given delivery ?

Most of these are complex constructs are essential in delivering rich experiences to customers, at the same time maintaining the velocity which is expected to keep users engaged. Codifying, automating, managing lifecycle, AI-driven intelligence of such ‘Application Intents’ is what delivers the key value for Edgeification.

Context

Doesn’t the whole cloud transformation lead to edgeification because not everything can instantly go or needs to go to the cloud. Hence, what’s left behind on-prem, does that become an edge?

Also, would it not be helpful to have a dashboard that gives you visibility to know that your Application Intents are being met?

--

--