Behind the scenes of smarter networks with our Director of Engineering, Monika Antoniak

To help us explore the upside potential of network automation, orchestration, and observability, we spoke with Monika Antoniak, Director of Engineering and Professional Services and network automation expert at CodiLime. With her vast experience in tackling challenges like scaling diverse infrastructures and integrating legacy systems, Monika shares her insights into how these technologies are reshaping network management for organizations of all sizes.

Hello Monika. Thank you for taking the time to chat with me today to talk about network automation. I think our readers are familiar with the basics, but perhaps they’d like to know how we should understand network automation in the context of modern network management.

Thanks Benjamin. It’s always great to discuss a topic so central to what we do at CodiLime. Let me start by saying network automation is essential because it addresses the challenges of scale and complexity. In modern networks, you’re often dealing with diverse equipment—hardware from different vendors, a mix of software solutions, both virtual machines and container-based systems, and more. Managing all of this manually, even with a skilled team, introduces errors and inefficiencies.

Network automation allows you to design repeatable processes that can be executed efficiently and error-free. For instance, once an automated workflow is set up, it can handle tasks like updates or deployments across thousands of devices with minimal intervention.

For example, in one of our recent projects, we automated the upgrade of 500 devices, completing the project in just three months. Given the constraints of maintenance windows, this would have taken over a year manually. With this type of approach, automation can reduce the time needed to complete the tasks by around 70%, and with knowledgeable teams, businesses can achieve even greater efficiencies—up to 86% cost reductions have been reported when automation is adopted.

That’s an extraordinary improvement and would obviously free up engineers for more business-critical tasks. We’ve also been hearing a lot about orchestration, and some of our readers are interested in learning more about it. How does it differ from automation, and what role does it play?

That’s right. The efficiency gains from automation cannot be overstated. Regarding orchestration, it is sometimes called the “automation of automation,” but it’s more than that. Automation focuses on specific tasks, like applying a patch or configuring a device. Orchestration takes a top-down view, focusing on the desired outcome or workflow.

For example, if you’re deploying services across two data centers, orchestration coordinates the underlying tasks—provisioning VMs, configuring networks, and so on—without requiring manual input at each step. It’s a user-centric approach: you click a button, and orchestration ensures the necessary tools and workflows are triggered to achieve the desired result. Of course, it is easier said than done, but that’s the idea.

This top-down approach contrasts with automation’s bottom-up focus on individual tasks. By combining both, you achieve end-to-end workflows that deliver real business value.

So its name is rather accurate; it’s like conducting a symphony of automated tasks like an orchestra. It’s clear how automation and orchestration work together to improve network operations and deliver real business value. But once these processes are in place, how about network observability? How does it improve on traditional monitoring and ensure everything runs smoothly?

That’s a great question. Observability builds on traditional monitoring but takes it a step further. Monitoring is reactive—you set thresholds, and if they’re crossed, you get an alert. Observability is proactive. It involves analyzing data across the entire environment to detect anomalies, even when they aren’t predefined.

For instance, with AI-powered observability, you can forecast bandwidth usage trends and make adjustments before congestion occurs. In one scenario, instead of waiting for a link to reach 80% utilization and then scrambling to add capacity, forecasting allows you to reroute traffic preemptively or plan infrastructure changes well in advance.

This approach is not possible without integrating data from various sources (and of various types quite often), creating a holistic view of the network. With forecasting and anomaly detection, well-deployed observability becomes a way to prevent downtime and maintain network health.

A proactive approach, rather than a reactive one, like having a crystal ball for your network to see into the future. It sounds beneficial for any business, not just those involved in network infrastructure. Can you tell me what are the key business drivers behind adopting network automation, orchestration, and observability?

Time-to-market and flexibility are the most important. Automation lets you launch new features or services faster, which is vital in today’s hyper-competitive market. It also allows for making processes more accessible to less specialized teams.

Flexibility is another driver. Automated systems allow businesses to adapt quickly to changes, whether that’s deploying infrastructure in new locations or scaling up to meet demand. This agility positively affects customer satisfaction and positions businesses for long-term success.

The benefits are clear, but I am sure introducing automation in legacy systems often presents challenges. What obstacles do organizations typically face?

Legacy systems are inherently complex. Many older devices lack modern interfaces like APIs, requiring custom solutions for integration. For example, you might need to develop scripts that use protocols like SSH to communicate with legacy hardware.

Skill gaps are another challenge. Teams accustomed to legacy systems often lack experience with newer automation tools, which can delay adoption. Observability faces similar issues—legacy systems may not provide the data needed for modern analysis or a way to get them, necessitating additional adapters or workarounds.

However, these challenges can be addressed with the right expertise and approach, which is why many of our customers choose us. At CodiLime, we specialize in tackling the complexities of legacy systems by designing custom solutions that integrate with modern automation tools. Our team has deep experience in developing adapters, building tailored workflows, and providing the necessary training to upskill internal teams.

Yes, I can see why choosing a partner with experience integrating network automation in legacy systems could be the better option. However, that option might involve more upfront costs for a company, and measuring ROI can be tricky. How can a business evaluate the impact of their investments in these technologies?

Yes, you’re right, but there are both direct and indirect metrics a business can use. Directly, you can measure faster deployments, reduced failures—especially those caused by human error—and quicker recovery times.

For instance, downtime costs can reach up to $9,000 per minute in some industries, and for 98% of organizations, one hour of downtime costs over $100,000. Faster recovery minimizes these losses and also helps maintain service continuity improving customer satisfaction.

Indirectly, automation frees teams to focus on innovation rather than maintenance. This shift drives business growth and long-term value.

That makes sense, so the upfront costs associated with implementation can be recovered pretty quickly. So, with that in mind, what advice would you give to businesses starting their journey with network automation, orchestration, and observability?

I would say start small and focus on areas with the highest impact. A step-by-step approach allows your teams to adapt and gain confidence in using these technologies.

Also, choose the right partner—someone with experience who can guide you through the process. This minimizes risks and ensures a smoother transition.

Great advice, Monika. Thank you for joining me today. I’m sure our readers will want to know more about network automation and how they can start implementing it in their own environments.

Thank you for the chance to share some knowledge. If your readers want to know more, there are many resources on the CodiLime website and YouTube channel where they can get to grips with the specifics.