In Networking, we’re all sick of hearing it… it generally goes something like this: “The network is the long pole in the tent, I can spin up VMs or go to the public cloud in minutes, but it takes days/weeks to get the network (i.e., VLANs/ACLs/Firewalls) configured”. While easy to complain, this is a difficult thing to fix, as it requires technology, process, and people changes.
Networks need to be more agile, we need to overcome network incrementalism, re-think how to manage risk/downtime, and make our networks more Googley. We’ve published on this topic several times, and we are now slapping a term on this much-needed shift. We are calling it (…wait for it…) Netops 2.0, in the hopes that putting a brand on it will increase awareness and accelerate the required investments in a Netops 2.0 direction.
You could argue that other terms capture this essence already, but I will argue they don’t. DevOps is too broad, while SDN is product-centric and means everything and nothing at the same time. Some of the key guiding principles behind a NetOps2.0 mindset include
- Networks must be Agile, not Fragile
- Automation by default, manual by exception
- Tools should be used proactively, not just reactively
- Risk should be managed, not avoided
- API > CLI
- That’s the way we’ve always done it can get you fired
There are several key steps in transitioning towards Netops2.0, including embracing network automation, using network analytics tools proactively (we have the technology), and then (longer-term) looking at Intent-based networking and algorithmic IT Ops. Here’s the link to the full research, led by my colleague Sanjit Ganguli (or, as spell check prefers to call him, Scanjet):
Summary: Today’s network operations teams will require a significant skills shift to take advantage of advances in network automation and analytics and to meet the demands of digital business needs. I&O leaders must champion a paradigm shift to NetOps 2.0 to stay relevant.