The key question is why network automation projects fail? I think it’s because we don’t look at the whole picture. We look at tools, people and budget separately, while they are all related and depending on each other. In this blog, I will zoom in on why automation projects fail and how you can succeed.
Reasons for failure
Let’s first look at the research done in this field by analyst group EMA. Below is a summary of the findings split into two categories: Tooling specific- and business-related reasons for failure.
Tooling specific reasons:
- Price - Simply said many of the tools are just too expensive. Companies are not able to deploy (commercial) solutions at an acceptable cost at scale.
- Difficulty of implementation - It’s a hard reality, network automation can be complex and daunting. Tools are not magically gonna solve this, whether they are commercial products or homegrown solutions.
- Scalability - It’s one thing to write a script for a specific task. But to deploy a solution at scale for all staff with adequate support is in a whole different ballpark.
- Breadth of automation - Customers have come to realize the diversity and complexity of use-cases. And uniqueness for each network. One single tool cannot get the whole job done unless you choose for a platform with enough flexibility and openness.
On the other side, EMA identified the following business-related challenges:
- Security risk - Companies are concerned that automation tools don’t meet their corporate security standards. This is especially the case for open source software and homegrown solutions where no external party is willing to cover these risks.
- Budget allocation - Network managers find it difficult to allocate budget to purchase network automation tools even though enough Opex is spent yearly on staff and running the network.
- Cultural resistance - Many network professionals struggle to visualize how solutions could help them and they are afraid that automation will break their network. Unless coached by automation experts they are likely to hang on to their old (manual) way of working.
- Skills - Although recognised by network managers as a priority, many day-to-day tasks and other priorities seem to overtake the need to develop automation skills.
In short, all these reasons can be summed up into three main categories: People, Tools and Budget. But what many fail to realize is that these categories cannot be seen separately. The key to success is to find a way to tackle all reasons for failure simultaneously.
The key to a successful network automation project
Let’s take a look at the different relations between people, tools and budget:
1. People & Tools: A tool is only a half product
People often make the mistake to think that tools are the answer to their network automation challenges. But a tool is only a half product, you need the right people to make it work. And here comes the next mistake: to think that you only need one type of skills.
For network automation tools to work, you need networking skills, knowledge of the customer’s specific network, experience with network automation and software development skills (when building in-house).
Ideally, you have DevNet specialists whose sole purpose is to build and maintain network automation solutions and train the customer’s engineers on-the-job as part of a continuous process. This will speed up any implementation, minimize cultural resistance and create acceptance throughout the organization. In the end, that’s what determines success.
At the same time, you also need people who support the system running the solution and who deal with software updates, upgrades, bug patches etc. Finally you need to make sure that this is not on one off but a continuous process.
2. Tools & Budget: Traditional ways don’t work anymore
The research shows that the biggest challenge with network automation tools is price. The tools are simply too expensive. So it seems logical to develop a solution in house with for instance, open source software. However, while you may think you save costs, many companies find out the hard way that there’s much more to building software and maintaining a system that complies with the company’s requirements. As the EMA findings also point out, security, breadth and scalability are the other reasons for failure.
When going for commercial tools, you need to allocate budgets and go through tedious selection processes. We still see companies spending an enormous amount of time and effort to shortlist and select vendors, only to come to the conclusion they do not exactly know what they actually want or need. The good old RFP process does not work anymore, and this is especially true for network automation. Tools are only half products, and you need automation experts and engineers with knowledge about your specific network to make it work. When you don’t have these experts ...Then there is the risk of selecting the wrong tool. Companies typically find out the hard way if they made the right choice: from their own hands-on experience. And experience always costs money ;-).
I think it’s crucial to first test a solution in your own environment. Not in the form of pilots or proof of concepts, but by implementing small business cases where you and the tool vendor share investments and risk.
3. People & Budget: Smartly reallocate existing budgets
One of the reasons for failure is cultural resistance. Many network professionals struggle to see the benefits of network automation. That’s why it is so important to have experienced automation experts help you transform this sentiment and train your current staff on a continuous basis.
A common pitfall is to send networking staff on a training course and have them learn network automation on the side, as a second skill. This might seem like a good idea, but in practice this is not enough. Their key responsibility is managing the network and this will always take precedence over automating the network. So where do you find these people with automation skills? Big companies, take Facebook or Google as an example, have the budgets and are really good at attracting and managing talent internally. But let’s be honest; when you don’t have the budgets or the name, you have to find more creative ways to succeed.
The thing is, available money is already spent on staff, contractors, suppliers, equipment and maintenance. My advice would be to first reallocate a part of this budget to select a (small) project that proves the business case. Or find a partner that is willing to share this risk with you.
We chose to adapt a business model that supports this, a.k.a. subscription or at-a-service models for Dev/Net skills. Feel free to reach out to firstname.lastname@example.org to learn more.
- When you want to succeed with network automation, you should look at tools, people and budget together since they are so dependent on each other.
- Investing in only one area will almost guarantee failure; rather spend your budget on both people and tools together than invest a large sum in one of them.
- The key is to start with a small project that proves the business case.
- You should invest in people with network automation skills (Dev/Net) and make sure they stick around, either do this yourself (but you need a budget) or find a partner that can provide this for you.
- Don’t underestimate the skills needed to support your network automation tools from an IT perspective.
Our #1 priority is helping you in your automation journey. Just schedule a meeting with me! I love to hear about your specific questions.