3 Data Center Problems Solved Using Zero Touch Provisioning

Zero touch provisioning (ZTP) is all about automating the provisioning and configuration of new equipment. ZTP uses DHCP to help scale the data center with fewer manual actions, simply plug it in and walk away. Hence the zero touch naming convention. 

We have chosen three data center problems solved by zero touch provisioning to help you understand why ZTP is so important. For the context of this article, we will detail how it works with Junos systems.

How Does ZTP Work?

ZTP makes use of DHCP, which automatically grabs a network configuration for a device that connects to the network. Simply plug in the Juniper device on either of its two DHCP interfaces (usually ME0/EM0 or IRB0) and the automated processes will take care of the rest. 

If any of these interfaces are plugged in, then the switch or any other Junos device that needs configuration can get an IP address from the pool. As you probably know, DHCP offers options like default gateway, DNS, NTP, and TLV (Type Length and Value).

TLVs are very useful because they allow you to define options that give your devices set instructions on what the next steps are during the provisioning stages of an installation. These options are things like the desired image for an operating system installation. 

The device then starts to pull that OS image down so that it can be flashed to the Junos device. It checks its version, and the TLV also checks the TFTP server and then runs through the upgrade process after downloading the image.

Once the OS is upgraded, TLVs can tell the device if there are any specific configurations available for that IP address.  After that is done, then the scripts that Ansible has configured can run and finish the setup. It is very similar to how telephones and wireless access points have worked for years, so it's not new technology, but rather is a relatively new implementation for network equipment. 

ZTP means that all you need to do is unbox, rack, cable, and plug in the device. ZTP will take care of the rest of it. You can concentrate on provisioning multiple devices and move on to the next one without waiting for them to successfully finish installing themselves. 

Let’s look at the main problems that ZTP hopes to solve.

1. Manual Labor

ZTP looks to do away with manual labor. Data centers get new hardware all the time, and there is a lot of manual labor with unboxing, configuring, and managing each new device.

You might be wondering what is wrong with manual labor, and the answer is nothing, in small doses. When you scale up monotonous tasks that are redundant and a bit of a waste of time then we simply have to look at automating as much of it as possible. 

Let's look at the below when installing a new piece of networking equipment in the data center as an example of manual labor and repetitive tasks.

Onboarding a new switch generally looks like this: 

• Travel to the data center

• Unbox the equipment

• Rack it (install it into the cabinet with cage nuts)

• Cable it

• Load the correct OS image from your USB thumb drive. (You did remember to bring it with you, right?)

• Configure the individual rules and configurations for the switch.

• Test

• Troubleshoot (If there are issues)

All of these steps are fine if you are installing a single switch. But imagine that you have to install 10, 20, or even 100 new switches with a limited amount of time, it would be almost impossible to do without additional staff to assist you. That is a lot of effort, time, and money.

ZTP solves this by automating the provisioning of the switch, so you can configure and deploy the switch remotely. It also allows you to upgrade the switch remotely as well. 

It's a great tool for any organization that has a large number of switches that need to be deployed. All you need is the necessary physical installation to take place, plug it in, and start installing the next one while it automatically installs and upgrades the software and configurations. 

2. Human Error

Networking is no simple matter in complex environments like data centers. Software defined networking takes traditional networking concepts and can complicate things, and small errors can mean big issues.

When there is human error involved in a manual configuration such as incorrect IP addresses, bad configurations, or incorrect operating system versions, then this can cause downtime and added time to troubleshoot.

Working on a scheduled deployment is critical, and when problems strike there are untold complications and added stress. Automating the whole setup takes these issues out of the equation and makes for a more streamlined project.

3. Difficulty at Scale 

The whole point of ZTP is that nothing needs to be configured, only the physical installation needs to be carried out. This is even more useful as we look at custom environments that need special configurations and scripting to get up and running.

After ZTP finishes with an OS image, then it looks for a configuration, and then it looks for automation from Ansible or your Python scripts. This is the magic sauce that makes automatic provisioning and rolling out new hardware plug-and-play. 

The image comes off of a TFTP server but that also needs configuration like IP, DNS, NTP, Root password, SNMP S/Flow, and NetConf. If you were manually deploying this equipment then at this stage you are finally able to get SSH running so that the actual configuration can start. 

This means that most of your time spent as an engineer on site would have traditionally been spent just to get a device online, and that was for just one device. Data centers can have hundreds of devices that need to be rolled out and this has to be done for each one.

If you multiply this operation and then quantify this into hours worked, then it is not feasible to have it done manually, and ZTP is essential for any deployment that requires a lot of equipment to be installed within a specified time frame. ZTP then completes the final steps while you continue on with other work.

Final Thoughts

There are many challenges that data centers face when trying to operate, deploy and scale. We have looked at three of the biggest and most common issues that you are likely to come across, and how ZTP solves them.

By leveraging this technology you can free up your time to focus on developing further automation and making existing systems more efficient and with even more features.

Networking has gotten to a level of complexity and scale which means that being able to write scripts and automate processes is the difference between getting the job done ahead of schedule, and not getting it done at all. ZTP helps get the job done — and learning how to use it will save you a lot of time, effort, and trouble.