Building an Edge Computing Strategy. Getting Started ( Part I)

Edge Computing is becoming a critical component, enabling organizations to leverage edge data in unprecedented ways. However, embarking on an Edge Computing project can be daunting, as building an edge solution requires significant time, investment, and a highly skilled team.This article explores the essentials for getting started with Edge Computing.

Industry at the Edge
Written by:
Miren Zabaleta

Getting Started with Edge Computing

In the industrial space in particular, Edge Computing is becoming instrumental for organizations looking to take full advantage of the Internet of Things (IoT) and other edge-oriented technologies. With the explosion of connected devices and the need for real-time data processing, it is no longer practical to send all data to a centralized data centre. Edge computing minimizes latency and bandwidth usage by reducing the distance data has to travel leading to faster decision-making.

However, despite the abundance of industry data collected in recent years, less than 25% of that data is ever processed. The main bottleneck? Many enterprises lack an edge computing strategy and a proper "build or buy" edge infrastructure to harness edge data effectively.

1. Identifying use cases and evaluating latency requirements

Companies need to determine the specific applications that will benefit from edge computing.

In manufacturing the most popular uses cases are predictive maintenance of machines, quality control and real-time monitoring of equipment. It is important also to understand the latency sensitivity of the applications to prioritize edge deployment where it matters most.

A use case well-suited for an edge computing solution will require one or more of the following:

• Low Latency: When rapid response is needed, or the response time needs to be deterministic and predictable, computing capability needs to be physically closer.

• Large Data Processing / Cost of Bandwidth is an issue: As data at the edge grows in volume, the cost of sending noisy, ephemeral data elsewhere to be filtered or processed might be more than the cost of moving the compute to the data.

• Limited Autonomy: Enterprises may need to maintain a working environment even when the connection to the central data center or cloud service goes down or is unavailable for some time.

• Privacy/Security: Enterprises may prefer to keep certain raw data locally, or the data might be regulated (e.g., facial recognition data). Regulatory requirements might also vary for different edge locations.

2. Infrastructure Planning

Edge Infrastructure: Enterprises should access to a robust edge infrastructure that includes edge servers, gateways, and networking components. This infrastructure should be scalable, resilient, and capable of handling the anticipated data load.

Connectivity: Companies need to ensure reliable and high-speed connectivity between edge devices and central systems. This can involve a combination of wired and wireless networks

3. Integration of AI and Model Management

Edge AI:  This involves deploying machine learning models on edge devices, allowing for immediate insights without the need to send data back to a central server.

Digital data production at the edge is growing exponentially, creating the opportunity for deeper analysis, automation, AI and ML. In 2022, perhaps 5% of edge computing deployments involve some level of ML — but by 2026, at least 50% of edge computing deployments will involve it.

While inference is usually implemented at the edge, training is often done by sending massive datasets to central processing. However, as the cost of compute continues to decline compared to the cost of bandwidth, training may also be done closer to the edge where the inference models are deployed.

Model Management: Develop a strategy for updating and maintaining AI models at the edge to ensure they remain accurate and effective.

Want to know more about Model Management and Deployment at the edge? we recommend the following MLops at the Edge

4. Security and Compliance

Data Security: Companies need to implement robust security measures to protect data at the edge. This includes encryption, access control, and regular security audits.

Compliance: Also ensure that your edge computing strategy complies with relevant regulations and standards, such as GDPR for data privacy.

Learn more about "Industrial Cybersecurity and Edge Computing"

Keep abreast with Edge Computing.  Register to our upcoming webinar.

Join us as we we'll explore the challenges faced by leading organizations in their digital journey, "from data collection to on-site analytics and real-time inferencing".

Building vs. Buying an Edge Infrastructure

Edge environments are highly complex and heterogeneous. If you are thinking to scale at speed, it’s worthwhile considering an edge computing platform that can support your growth. The task of managing edge operations across diverse locations, devices and applications with the highest security standards can be daunting and expensive.

When it comes to integrating and Edge Computing infrastructure within the digital transformation plan,  organisations face a pivotal decision: should they build their own custom solution or buy a third-party offering?

Buying from a Third Party


1. Speed to Market: Purchasing a third-party solution can significantly accelerate deployment time, allowing businesses to benefit from Edge AI capabilities more rapidly compared to developing a system in-house.

2. Reduced Initial Investment: Building an Edge AI infrastructure requires a substantial upfront investment in research, development, and testing. Buying a solution can lower these initial costs.

3. Expert Support: Vendors often provide ongoing support and maintenance, ensuring the system remains up-to-date and operates efficiently without requiring in-house expertise.

4. Proven Solutions: Third-party products have typically been tested and validated across multiple deployments, offering a level of reliability and performance assurance.


1. Less Customisation: Off-the-shelf solutions may not fit every unique operational requirement, potentially leading to compromises in functionality or performance.

2. Ongoing Costs: While the initial investment might be lower, recurring licensing fees, subscriptions, or service charges can add up, impacting long-term budgets.

Building Your Own Edge Infrastructure


1. Customisation: Building in-house allows for bespoke solutions tailored precisely to an organisation's specific needs, offering optimal integration with existing systems and processes.

2. Control and Independence: Owning the infrastructure reduces dependency on external vendors, providing more control over the technology stack, data security, and future developments.


1. Higher Initial Costs: The costs associated with research, development, and deployment of a custom solution can be significantly higher, requiring substantial initial investment.

2. Longer Deployment Time: Designing and building a bespoke system is time-consuming, potentially delaying the realisation of benefits from Edge AI.

3. Maintenance and Support: Organisations must allocate resources for ongoing maintenance, updates, and troubleshooting, requiring in-house expertise or external consultants.

Barbara, The Cybersecure Industrial Edge Platform

Barbara Industrial Edge Platform helps organizations simplify and accelerate their Edge App deployments, building, orchestrating and maintaining easily container-based or native applications across thousands of distributed edge devices.

  1. Real-time data processing: Barbara allows for real-time data processing at the edge, which can lead to improved operational efficiency and cost savings. By processing data at the edge, organizations can reduce the amount of data that needs to be transmitted to the cloud, resulting in faster response times and reduced latency.
  2. Improved scalability: Barbara provides the ability to scale up or down depending on the organization´s needs which can be beneficial for industrial processes that have varying levels of demand.
  3. Enhanced security: Barbara offers robust security features to ensure that data is protected at all times. This is especially important for industrial processes that deal with sensitive information.
  4. Flexibility: Barbara is a flexible platform that can be customized to meet the specific needs of an organization. This allows organizations to tailor the platform to their specific use case, which can lead to improved efficiency and cost savings.
  5. Remote management: Barbara allows for remote management and control of edge devices, applications and data, enabling organizations to manage their infrastructure from a centralized location.
  6. Integration: Barbara can integrate with existing systems and platforms, allowing organizations to leverage their existing investments and improve efficiency.