Why edge computing matters to enterprise software

Edge computing means processing data near where it's produced and consumed. It's another dimension to that age-old workload placement question, "Where's the best place to do this?" The trend this century has been to move computing into the data center or the cloud — now we have to consider the edge as well.

In many ways, edge computing is reminiscent of the client-server model of the 80's and 90's in that some tasks are performed closer to where the outcome is needed. But it differs in that it takes place on a generation of more powerful devices that didn't previously exist, like smart phones.

Interest in edge was originally driven by a desire to reduce latency and bandwidth. Today the main drivers are the internet of things and real-time apps that demand instantaneous data processing. 5G is accelerating this trend, with use cases in process control, optical inspection, autonomous vehicles, and augmented reality experiences.

What does edge mean for enterprise software?

Edge has a different meaning when it comes to enterprise software. Here it's not primarily about real-time processing — that's not a common requirement of enterprise resource planning (ERP), financial planning and analysis, or human capital management systems.

In the enterprise software world, edge is more about distributing pieces of a previously monolithic application out to the people who use it. Workers don't need access to the whole application suite and all corporate data. They only need part of the application and the data necessary to perform their specific function.

Only giving people what they need, where they need it, improves response times and makes for a better user experience. Putting a task-specific app with a limited data set on the user's own device, for example, will make it easier for them to get their work done.

Four types of edge in enterprise software

There are four ways we can interpret edge for enterprise software:

1. Architectural edge — This is about what kind of system is being distributed; front-office systems can be placed at the edge while systems of record are held centrally.
2. Geographical edge — A global company might have its systems of record in a data center in the HQ country, but place subsets of data in local countries to improve response times or comply with regulation.
3. Infrastructure edge — In this sense, edge equals distributed cloud. Even edge apps will have to run in some sort of scalable environment.
4. Application edge — Instead of a user having to access the whole system, you create task-specific apps in which a small part of the application is hived off and provided to the user along with the data they need to do their job.

What are the business benefits of doing this?

First and foremost, it improves the People Experience — making the app easier and faster to use, more mobile- and homeworker-friendly.

Next, it enables you to harness the potential of AI. Workloads like voice recognition need to work locally for fast response (even in ERP), but the algorithm can be trained in the cloud. A distributed application architecture makes compliance with data sovereignty regulations easier to manage too.

Overall, distributing elements of enterprise software out to the edge makes for easier, faster and more accurate work. And it's a lean, agile approach more readily adaptable to changing circumstances.

The capabilities you need to be able to do it

First of all, moving enterprise software towards the edge requires a passion to continually improve People Experience — the desire to make work easier by making software easier to use.

Your enterprise applications will need to have a microservices architecture, with processes that are event-driven and message-based so that transactions can be distributed.

Next, you need the ability to construct front-office systems as separate units within the microservices architecture, relating to central systems of record on a trust-but-verify basis.

Finally, you need a watertight means of ensuring transactional integrity and data fidelity across many locations to be sure that data at the edge and in the core are aligned. Although hyperscalers can guarantee data delivery, you need separate tools to make sure data is assembled in the right order.

Edge is very relevant to enterprise software — just in a slightly different way to how it's normally understood. In essence, it brings power closer to the user. So, when you're considering the future of your enterprise software, don't forget to ask your vendor about its edge strategy.