How do you select network automation and scripting tools?

The ability to run detailed reports on problem/trouble ticket/incident types is helpful in determining what use cases will have the greatest returns. Make sure you are properly categorizing these events, and can report on them to help choose which tasks to automate.

For our specific use case repetitive daily ticket issues have forced use scripting and automation to help resolve this issues. - Clearing err-disabled ports (and adding logic to exclude certain conditions and instead alert the tech) - Clearing forcing re-auth of ports (correcting failed MAB or 802.1x auth attempts) - Configuration backups and change auditing

Do you do repetitive tasks, especially ones that take valuable time? Regardless, of whether you have a small network or a large one, those repetitive tasks can benefit from automation or scripting. In a small scripting, may reduce costs. In a large network look at what tools you will do the automation with, do they have REST APIs; if any of them do not then that will impact your choice on tools. Do you need robotic process automation (RPA) or something more specific like cybersecurity automation? While a general RPA tool can fit many scenarios, they also take a lot more effort and resources to setup; however, a targeted automation tool may be able to be deployed very rapidly and see benefits almost immediately.

Determine why you need network automation & if it's the right path for you. For eg, a small network with a few devices, don't need it. If you have a large network network automation can speed up configuration and deployment, and automate mitigation of events. It gives you more control over your network and make it more agile Network automation has various use cases, and it cannot all be implemented overnight. Plan ahead and determine which type of automation you want: device provisioning, configuration management, migration, troubleshooting, or event mitigation Start with a simple CI/CD pipeline and gradually add more features. The goal is to have a fully automated pipeline that can test and deploy configuration and network changes.

Periksa semua konfigurasi, Cek semua peralatan pendukung. Restart system. Bangun jaring sesuai kapasitas dan kebutuhan. Setting automatic data dumming. Konfigurasi database security system 4 layer.

Research the devices on your network to make sure that you understand all the different methods that can be used to manage those devices. Many devices that are typically managed from a web interface (which makes sense at a small scale) can also be managed by SSH, etc.

Look at how you will interact, from an automation perspective, with the devices on your network. Will it be via API or something else like SSH. Do the commands you want to automate exist in the command set or data dictionary for the interface? I've encountered cases where an obvious command regularly used in the GUI interface, just simply doesn't exist in the REST API, so then you have to decide "do I work around this function, or do I use a tool that can make use of the other interface"? In my case it was a GUI tool that would be needed, which is also, potentially more fragile because any future changes to the GUI layout could impact the automation.

Typically , vendor agnostic approach like ansible etc shall help here. Also, Sticking to basic mechanisms like SSH , Telnet, APIs, SNMP, TR069 etc shall be very helpful.

We have a largely brownfield environment of devices that require a multitude of logic to perform automation at scale. Thus we leverage vendor agnostic automation frameworks wherever possible (such as Ansible, Netmiko, and NAPALM).

Each vendor device has different types of south bound integration capabilities. Some devices do not allow you to access them directly, but instead require you to execute and control them through the vendor's specific centralized SDN controller. It is important to read through the vendor's API documentation and supported management protocols to understand how to integrate with their devices. Most vendors support RESTful APIs and NETCONF. For examples: Cisco devices can be integrated with a SDN controller using the Cisco APIC-EM API. This API allows to manage Cisco devices, including configuring their interfaces, creating & modifying VLANs. Similarly, Juniper devices can be integrated with a SDN controller using the Juniper Contrail API.

It's important to balance your own skills with the expertise of your team. Take everyone's coding experience into account when creating a new solution. Prioritize tools and languages that are established. Newer tools might lack support and reliability. Also aim for fail-safe coding practices to ensure your scripts are robust. This approach should match your current abilities and support team growth by avoiding complex or unsupported technologies.

Some of the skills, training, languages, and/or tools that I think are helpful for people getting started on their network automation journey include: SNMP, git, CCNA DevNet, Python, Ansible, Netmiko, bash, Go, REST API.

I feel that having spent the majority of my career as a Systems Administrator has prepared me for the automation and scripting side of Network Administration. The goals are the same, "Automate as much of the repetitive tasks as possible to free up time to spend on innovation and future planning". By approaching each problem and breaking them down each into easily digestible pieces makes creating a workflow for the automation or scripting much simpler. My preferred programming language is python and frameworks vary depending on function. My future goals is creating some automated unit testing for a QA environment to validate changes and confirm functionality.

Traditional network engineers are not proficient in coding, and many of them avoid it altogether. However, coding is essential in today's world. At the very least, a basic level of Python scripting is required so that they can use NAPALM or Netmiko. The skills of a Network DevOps engineer is required. This includes understanding API calls, various data formats like JSON and XML, how Netconf works, and how to use Git for source control. Have a basic understanding of Linux operations and be familiar with the Linux shell. Start small by familiarizing the team with Python so that they can run basic deployment commands using libraries. They should also be taught how to use existing templating engines like Jinja and tools like Ansible.

It's important to look at the skill-set of those that will be implementing the automation and/or scripting tools. Do they need to know a specific language? In the world of cybersecurity almost all roads lead to Python, so regardless of whether you buy a "no-code" tool or the tool requires coding, Python shows up over and over again. There are good courses on Python coding that can be found on everything from YouTube to LinkedIn Learning.

If the network size is small then building something quickly and ensuring itvis maintainable shall help than spending time in evaluating different automation tools and spending relatively significant money on purchasing them.

For our space, budget is always an issue. This is why we largely rely on opensource technology and tooling. The most expensive item is our people. When automation/scripting are created they must be understood and maintained to remain effective for the organization. This adds the requirement for future hires to have a programming/scripting familiarity and narrows the hiring pool.

The main resources required are the servers for deployment and the employees who will manage them. If your employees are well-trained and you have the budget to deploy servers on-premises or in the cloud, then you should use open-source automation tools. You can run your entire automation end-to-end using only freely available open-source tools. Most of the tools have good community support, and the community is helpful. If you are concerned about tool support, you can purchase paid versions.

Budgeting for a tool is quite difficult - some tools can be in the tens of thousands of dollars, whereas others are in the hundreds of thousands of dollars. So which one do you need? Consider running a PoC with a limited set of use-cases to use as a "bake-off" between the products. This will give you a good expectation around potential licensing costs, as well as resource requirements for the implementation of each "bot". If one tool takes 40 hrs to implement a bot and another takes 400 then that's going to have significant impacts on your costing model and resource allocation.

When it comes to networking, it is also important to realize how the hyperscalers have shifted from global vendors like Cisco and others to making a switch towards TIP and OCP. Essentially, they asked themselves what features we need and what we don't and then resolving not to pay for stuff they did not need or want. This helped them massively save on capex and leverage cost effective alternatives.

Most of these steps are performed manually, which can lead to errors and are time-consuming as well. Businesses can save time, reduce mistakes, lower operating costs, and improve their networks by automating them.

All hyperscalers run their day-to-day network operations smoothly with less human involvement. They have network bots that monitor the network and mitigate events. Devices are configured and managed using CI/CD pipelines, which tests the configurations, devices and then automatically deploys them during off-peak hours. However, if not handled correctly, this can lead to outages, as was the case in 2021 when Facebook experienced a backbone outage due to a misconfiguration in automation.

It can be daunting when you first start down the path of automation/scripting. Let me say that the first step should be creating or integrating with a source of truth (example Netbox or Ansible inventory file). This will greatly help as you progress and it doesn't have to perfect to begin testing/developing. Also you can use virtual machines and containers to simulate network gear or servers for automation/scripting development. Godspeed on your automation journey!

Network automation refers to the process of automating tasks inside a network infrastructure, such as change management, configuration management, device discovery and inventory, device provisioning, etc.