Traditional Legacy Network Management

Traditional networking

For better and for worse, not a lot has changed about networks over the past 30 years. Since the Open Systems Interconnection (OSI) model was published in 1984, there has been a rise in function virtualization, a move towards bigger and faster boxes, and improvements in bandwidth. However, the configuration and operation of networks has become a predictable and tedious task. Once administrators have learned to operate a device, along with the fundamentals of networking and the OSI model, there is little further progress. While the size and scale of modern networks have exploded, few tools have emerged to operate and configure network devices at scale. Connecting to the Command Line Interface (CLI) and applying configurations manually- device-by-device, line-by-line- regardless of scale has long been the only methodology available to network administrators.

Modern network management systems do exist as appliances or specialized software. However, many do not offer much beyond a graphical user interface (GUI) representing the line-by-line commands. Even more time-consuming is ensuring that network documentation is up to date- if it exists at all. In some cases, a real source of truth about the network may not even exist. The network running configurations device-to-device may itself be the source of truth.

This method of running a network increases the chance of human errors. Errors can result from clumsy fingers, bad copy-paste jobs, mistakes in the order of operations, or even using the wrong device! As long as we rely on manual changes, mistakes can occur. Network automation, however, minimizes such events. It allows teams to focus on higher-level tasks while using the automation engine to perform functions that are prone to human error. However, administrators will still make valuable contributions to the operation of the network. With the tools introduced in this course, operators can augment their abilities with the powerful automation engine- and do what was previously impossible!

A typical network change

Let’s consider a typical network change workflow an administrator may complete:

• They assess the new project’s network requirements, or changes that are required to the network.
• The administrator collects the data- possibly from multiple devices, often manually- The data lets them assess the current network state and draft the change.
• The administrator may first need to develop the change and test it in a non-production environment. Or, if the change is disruptive, they may need to perform an impact assessment.
• They then develop configuration commands.
• Once tested and ready, the administrator submits the changes into a ticketing system and assigns the tickets to a network operator.
• The operator reviews the change artifacts, and implements the change manually by following a series of instructions.
• Finally, the operator gathers the pre and post-change information to validate the change. This often takes a substantial amount of time, and will be point-in-time information.

The benefits of automation

If things do not go well, or mistakes happen, it is often difficult to identify the root cause or confirm if the operator followed the correct steps. Problems caused by a change are usually discovered because of an outage. Ideally, there are network monitoring system alarms that are triggered. But in a worst-case scenario, it is a user who reports the issue. To fix the problem, the administrator will need to manually troubleshoot the problem and updates all relevant documentation. Network automation solves all these problems while drastically reducing the time and effort involved in gathering information, resolving problems, and deploying changes to the network.

Aside from safe, pre-approved changes, the majority of network configuration changes occur after business hours. Several network administrators may be required to perform changes manually at a larger scale, or a smaller team may need a longer change window. By automating solutions, the organization can dramatically reduce outage windows, as well as the work hours needed to implement changes. Changes that previously took hours to be executed can now be performed in seconds.

The drawbacks of tradition

The drawbacks and problems inherited by tools such as console cables. Telnet, SSH sessions, and keyboards have impeded the progress of modern networks, especially at scale. PuTTY session copy-paste only goes so far.

A lot of Network Management Systems (NMS) have emerged from both hardware vendors and third-party companies trying to meet the need for better network management tools.

However, due to the huge number of requirements, an NMS must offer:

• Monitoring
• Reporting
• Configuration management
• Provisioning Capacity
• Utilization
• Performance statistics

Large NMS tend to do most things well but- nothing great.

Intent-based solutions are only starting to emerge as appliances or software and are only capable of delivering solutions for a single vendor. NMS solutions also involve licensing and can become costly as the network scales. Most NMS require training to be fully leveraged, and are often underutilized by an organization. This is either due to a lack of the NMS’s ability to provide solutions or the complexity of building solutions in the NMS.