Javier Solis

3 minute read

Being a network administrator/engineer typically requires typing in ssh consoles to get things going. At some point, being able to automate tasks or being able to manipulate configurations based on a certain outcome will become necessary. I’ve gathered a few thoughts on real world views to network automation. The buzzword floating around for this topic is NetDevOps.

NetOps/NetDevOps(my definition): Network automation using code to run commands that would normally have to be typed in manually into each device. Example: Run code that can parse or write through configs, logs, and snmp values in order to take action on a specific outcome.

I won’t go into the details of the ins and outs of NetOps/NetDevOps and how to get started with coding. I’ve provided a list of links with information that other really smart people came up with.

Detailed definitions:



Some examples:


Getting Started with coding


Ok, now what can NetDevOps actually do for you network administrators out there? I started to create a list of items that NetDevOps could put a dent in. I don’t feel that you require a Google or Facebook sized infrastructure to take advantage of NetDevOps. My team and I currently manage around 120 switching/routing devices and we’re headed to add lots more. That’s no Google, so here’s my list:

  • Changing network admin passwords for devices when staffing changes can be very time consuming. Running some code that can SSH into switches and routers to update passwords and privileges could be a very useful feature to have.
  • If you have routers that maintain the same ACLs or route maps, having to make changes can be a daunting task even if it’s only a dozen routers. Using code to automatically upload new changes to duplicate ACL’s and route maps across your routers will reduce time and human input errors.
  • I also have a few ideas about gathering wireless user data and plotting details on a google map to indicate AP’s that have a large volume of user connectivity. The map would provide visual information in real time that can help determine if you’re having a sticky client situation. You could also make some automated config changes to your AP’s power levels or with minimum basic rates to help the situation out.
  • You could manage bandwidth available across network links and trigger an automatic response to apply route maps to redirect traffic or apply QOS rules to your programming hearts content.
I’m sure there’s lots of other examples out there, please post others that you may have. I know this last one is heading down the openflow rabbit hole, but hey if you could do these types of things with your current equipment using a NetDevOps approach, why not?