You’re told to automate the network. Now what?
Enterprise Management Associates research reveals what steps you can take now to successfully adopt network automation (hint: it’s not learning Python).
Network automation is a high priority—learn to take the right steps
Network automation is a high priority for IT organizations. In fact, after surveying 250 subject matter experts, Enterprise Management Associates (EMA) research has found that 90% of enterprises are expanding their use of network automation.
Most recently, EMA’s research revealed that a network automaton initiative drives 43% of network management strategies.
But if you’ve been told to automate your network, what, exactly, should you do? EMA has found some key trends, perhaps giving you some food for thought for your next steps. In particular, this post will cover:
- First, network automation approaches vary widely and are mostly a multi-solution endeavor.
- Second, automation is hard. It is not without challenges.
- Finally, there are specific steps that organizations are taking–and networking professionals can take–to adopt automation successfully. (Here’s a preview: the answer is not learning Python).
P.S. We’re talking a lot about this in the Network VIP community on Slack. Join us!
Network automation is a diverse, multi-solution endeavor
While it’s clear that IT organizations are prioritizing network automation, there is less consensus on how to proceed. After all, network automation is not a new technology. Network change and configuration management (NCCM) tools have provided automation around version control and large-scale change management for years.
Network engineers have also built large libraries of one-off scripts and screen-scrapes that can automate simple, repeatable tasks. However, NCCM solutions have a limited scope, and one-off scripts are brittle. A simple switch OS update can break them.
Today, IT organizations have become more ambitious with automation, using a variety of commercial and open-source tools that can provision and configure infrastructure, orchestrate end-to-end services, automate problem isolation and root-cause analysis, and even support closed-loop operations.
Approaches vary widely
EMA research has found that enterprises vary widely in how they approach automation. It depends on their automation goals, their available resources, and their vendor partnerships.
For instance, one project might focus on data center network automation. Another might focus on the local- or wide-area network. Very few solutions are going to cover every aspect of networking from the data center out to the user edge.
Multi-solution is the norm
In most cases, network automation is a multi-solution endeavor. EMA research has found that only 9% of network automation initiatives involve a lone automation tool.
Instead, 40% encompass two automation tools and 34% involve three tools. Another 9% have five or more tools. These projects involve network automation software that can support multiple infrastructure vendors, IT automation software adapted to networking, and automation capabilities embedded in an infrastructure solution.
The challenges to automating the network
Most enterprises have some kind of network automation in place today, but no one has automated everything.
Automation is hard.
For instance, companies choose the networks they build for performance, scalability, and security. Only after building a network do many IT organizations start thinking about automation. Thus, enterprises must often automate networks that were built without a proper foundation. Automation has to be bolted on.
Certainly, this explains why cost is a major challenge for network automation. The tools are often too expensive because they are developed to solve very complex problems.
Enterprises also struggle with the difficulty of implementing automation, the scalability of available tools, and the breadth of capabilities.
Data is also a roadblock. Before network automation tools make a change, they must understand the intent of the network and the current state of the network. This requires a source of network truth with a comprehensive collection of data. Assembling this source of truth can be a major project.
Only 26% of network automation initiatives have a single, authoritative database. The rest federate data from multiple data repositories. And many struggle to eliminate data conflicts in this source of truth.
Amid the automation rush, what should networking personnel do?
EMA research has found that 83% of companies are devoting resources to training their people for automation. And contrary to what you might hear, learning Python and other programming languages is a low priority for companies.
Instead, IT organizations want networking personnel to learn how to use the network automation solutions they have adopted.
This may require training from a vendor if a company has adopted a commercial solution. Data management and analytics become important drivers of automation, so some enterprises are encouraging their network teams to acquire skills around data science.
The importance of cross-team expertise
Furthermore, IT organizations emphasize the importance of cross-team expertise. Their networking personnel need to understand the tools and processes of other teams in the IT organization. Automation enables the network team to work collaboratively with other groups and increase overall IT agility, and the stats bear this out. For instance:
- Nearly half of network automation solutions are integrated with an IT service management platform.
- More than one-third are integrated with a security monitoring tool solution.
- Forty percent are integrated with a private cloud orchestrator.
The return on investment
IT organizations that successfully execute on automation can expect a number of benefits. EMA has found that security and compliance risk reduction is a top opportunity, followed by reduced human error, improved proactive problem prevention, and the introduction of self-service infrastructure.
Also, 91% of network automation initiatives improve a network team’s ability to address change requests, which will be music to the ears of most DevOps teams.
So, what can I do?
With all this in mind, if you are facing a network automation mandate, don’t think you can ready yourself by learning Python. Instead, you should:
- Research the network automation solution market and understand how those solutions might support your unique environment.
- Identify how automation would best help your organization, whether it be security risk reduction or change management efficiency.
- Clean up your network data so that you can have a single source of truth for your future automation pipeline.
- Prepare yourself to be a partner to your business through automation.
- Learn how the rest of IT works and how automation can support a cross-domain approach to IT service delivery.
Read more about EMA’s findings on DDI solutions and a host of other networking trends in Network Management Megatrends 2020: Enterprises Embrace NetSecOps, the Internet of Things, and Streaming Network Telemetry. This report includes a case study with the University of Pittsburgh, which saw improved networked resiliency and control with a new BlueCat-based network architecture.