Network automation enables teams to use software to plan, develop, operate, or optimize networks with little or no human intervention. Effectively, network automation leverages some logic to execute “task A” when “event B” happens.
Network automation can be used in a range of ways, anything from AI-driven network analytics to traditional health checks. Do you want to know how network automation can help your business? Keep reading.
Why is network automation important?
If you ask network leaders what their network automation drivers are, you will normally hear one of these answers:
- Reduce cost. When engineers spend significant amounts of time on network changes or troubleshooting, it can have a huge impact on operating expenses.
- Reduce risk. Manual changes are often the cause of network downtime.
- Increase value. When teams are freed from manual operating tasks, they can invest more in innovation.
Here’s one real-world example of how network automation can help: The team at Cisco employed network automation to dramatically reduce the provisioning time of network functions. In the process, they reduced user errors and improved customer experience.
In a more basic scenario, network automation is used to manage everyday network tasks and repetitive processes. However, it is also used for more powerful use cases, including in conjunction with network virtualization technologies.
How is network automation employed in network virtualization environments?
Given the increasing implementation of network automation in virtualized environments, it is important to understand how these virtualization technologies work. Here are some of the most common approaches:
- Intent-based networking (IBN). This is an approach in which network administrators set the desired outcome or intent, such as “enable any IoT device connected to this network.” Based on the intent specified, the IBN system automatically triggers the associated execution steps, for example, configuring required ACL, VLAN, or QoS policies.
- Network functions virtualization (NFV). This technology is used to run network functions or components, including firewalls and load balancers, in standard hardware. NFV provides great flexibility and reduces cost. Security has been one of the first key areas to benefit from NFV, enabling the virtualization of firewalls that traditionally run in dedicated hardware.
- Software-defined networking (SDN). This network management approach segregates the control plane (controller or orchestrator) from the data plane (network devices forwarding the traffic). For instance, hyperscalers use SDN to dynamically allocate network resources across a pool of servers, based on shifting workload demands. This technology offers agility, efficiency, and scalability.
Users interact with these virtualization technologies via network automation tools, such as Red Hat Ansible, NetBrain, or VMware NSX. These tools interface with SDN, NFV, and IBN and enable the user to “program” or “manage” automated tasks or playbooks as a response to an event. These tools support one or more network virtualization technologies, including those from such vendors as Cisco, Arista, Juniper, and more.
Let’s dive into specific use cases and see how you can benefit from them.
What are some common network automation use cases?
Depending on your organization’s maturity and industry sector, you may realize value through a number of use cases. Here’s a list of common use cases, from simple, routine tasks to closed-loop remediation:
- Network monitoring. You can automate monitoring tasks, including health checks of problematic devices, traffic analysis, port scan reports, or enrichment of an alarm before creating a ticket. These are usually read-only, non-intrusive actions, which means they can be considered low-hanging fruit. Automating these tasks can add value for any organization, enabling teams to improve mean-time-to-resolution (MTTR) metrics.
- Configuration management. By automating the configuration of devices, teams can scale their operations while reducing manual configuration errors. These configuration tasks can also be triggered when non-compliant configurations are detected, helping to ensure consistency across the network. Virtually any team managing a network can benefit from this use case by reducing risk and cost.
- Security. Teams can run vulnerability scans of compromised devices or networks, enforce security policies, or trigger mitigation rules in the event of a cyberattack. For example, you can block traffic to prevent a distributed denial of service (DDOS) attack. This is a foundational use case and should be aligned with the company’s security and compliance teams.
- Provisioning. This approach can reduce the time required to deploy new services. This is particularly useful for organizations leveraging cloud resources or managing very dynamic and elastic networks.
- Optimization. Teams can improve network efficiency and mitigate congestion by automating route selection, QoS configuration, and load balancing. This use case is common to SD-WAN providers but it can benefit any communications service provider.
While the use cases above are not new, they are tremendously useful for improving network efficiency, reliability, and security. Here are several more innovative use cases:
- Self-healing. Through this approach, teams can detect and resolve network issues, without human intervention. This is achieved by automating the reconfiguration of network devices in response to changes in network conditions. For instance, a degradation in the baseline of key network metrics (jitter, packet loss, or latency) can trigger a re-routing of traffic. This use case can be seen as the holy-grail for many organizations. Currently, only a few specific vendors provide this capability and achieving cross-vendor self-healing is extremely difficult.
- ChatOps. Teams can employ a conversational interface that enables operators to interact with tools, systems, and workflows. This can help improve collaboration and efficiency.
- Orchestration. Through this approach, teams can automate end-to-end processes like provisioning a new resource, deploying security policies, and setting up and operating monitoring. This approach can yield direct improvements in service delivery speed and is a very relevant use case for organizations implementing DevSecOps.
- AI-driven network analytics. This enables teams to gather and analyze big data from networks. By using machine learning algorithms, teams can uncover patterns and anomalies and predict network issues. This is an ideal approach for proactively managing large networks and extracting actionable insights from the sea of data that SDN and NFV generate.
What are best practices for getting started with network automation?
Implementing network automation can be a big effort at the beginning but it will start paying off very fast. Here are some basic tips to get started:
- Start small. The “cheapest” network automation example would be to run an automatic knowledge base search as a response to an event and enrich the alarm with this information so the NOC improves their MTTR.
- Implement automation with a security mindset. It is important to audit every execution, implement ACL according to least-privilege principles, and use read-only tasks whenever possible.
- Establish baselines. Validate network automation changes by baselining user experience before, during, and after the change. In this space, tools like NetOps can be extremely useful.
- Retain some control. Implement an approval process in the first stages of automation. This enables your team to build confidence and retain a degree of control, before letting automation run in a fully autonomous fashion.
- Consolidate automation tools. Having multiple Perl, shell, or Python scripts will add complexity as you will end up with a hard-wired solution to a specific problem on a specific device.
- Ensure your team has the right skill sets. Network complexity is overwhelming these days so invest in training and roles covering new network technologies.
Putting it all together
Network automation is a rapidly evolving field, and the number and type of innovative use cases available continues to expand. Every use case won’t be suitable for every organization. By investing in the right automation approaches and technologies, pretty much any organization can realize significant benefits, including improving network performance, facilitating transformation initiatives, and reducing cost.
I hope this post has helped to clarify what network automation can do for you. And, remember, when you put network automation to work, you can free up more time to read articles like this and deliver more value through innovation.