After more than two decades of effort, millions of man-hours, and hundreds of tools created, network automation has largely not achieved its intended objective of eliminating manual operations. Multiple industry association surveys (e.g., Enterprise Management Association, Gartner) show that most enterprises have automated less than half of their data center tasks.
Why is this so?
When asked, most IT network operators would say that “automation” should help them do manual, repetitive tasks faster. But fast alone is not sufficient. Doing things faster without including reliability in the automation operation does not move network operations forward. As Bill Gates said: “The first rule of any technology used in a business is that automation applied to an efficient operation will magnify efficiency. The second is that automation applied to an inefficient operation will magnify inefficiency.”
There are two primary reasons why network automation isn’t widely adopted: tools and people. Part of the blame belongs with the vendor community where there has been insufficient emphasis on operational automation, resulting in few solutions conducive to scalable operations across vendors. Tools are often cumbersome to operate and lack a fully unified machine interface, which hinders integration and management of diverse network elements. There is also a skills shortage. Certified network operations engineers are retiring, and new hires are more familiar with cloud network operations than traditional network CLI commands. The result for network operations is a lack of trust in the automation platform. “Is it going to work as expected?” “If not, is there a rollback feature?” “What if it causes an outage?” This lack of confidence in the automation outcome results in slow to no adoption and status quo for network operations.
Both trust and reliability are what inhibits all other efforts around programmability and automation. If reliability is not ensured, then any tools that make things faster are not addressing the root problem.
Originally, Kubernetes was designed as an open source, cloud community platform for container orchestration of infrastructure resources that are distributed across data center servers. Over the past decade it has evolved. Now Kubernetes is widely adopted by industry as a comprehensive and well-designed orchestration and automation platform. An abstraction layer and the use of a wide range of open-source tools optimize resources that are running on premises or in public or private clouds. These support a declarative configuration, which enables the automation of manual operations and processes and is backed by a large ecosystem of tools, services and community support.
Many Kubernetes attributes can be applied to network operations automation:
With a massive ecosystem of tools, plug-ins and extensions available, Kubernetes is the de facto standard as an automation and orchestration system. It is used across multiple domains such as containers, VMs, cloud networks, bare metal or storage systems, and server workloads.
If Kubernetes can be successfully used as an automation and orchestration platform for all these different domains and applications, then why not extend it to support automation and orchestration for network operations?
Event-driven Automation (EDA) is a modern infrastructure automation platform from Nokia that’s built on a Kubernetes technical foundation to bring highly reliable, simplified and adaptable life-cycle management to the data center network needs of the AI era. Its aim is to reduce network disruptions and service downtime by driving human errors in the network down to zero while also decreasing operational effort by up to 40%. Features such as digital twin, change rollback, version control, closed-loop automation and declarative intent-based automation increase reliability and trust.
A single dashboard for all operations simplifies life-cycle management, and Gen AI assistance, streaming telemetry with fine-grained visibility and proactive notifications ease operations and improve observability. EDA is highly adaptable with seamless integration into ISTM/workflow/trouble ticket systems, event notifications and cloud management platforms. It also supports a wide Kubernetes toolchain ecosystem and a multi-X operating environment (multi-domain, multi-vendor, and workload extension to multi-cloud). The EDA network automation platform is part of Nokia’s data center network solution which supports the extensive portfolio of Nokia data center switching and routing platforms that run the company’s Service Router (SR) Linux network operating system. With EDA, enterprise organizations can finally achieve what they have been missing: the ability to confidently move at the speed of business by leveraging a modern, automation platform – knowing any changes made by the system can be trusted.
For more information, read our latest press release.
If you would like to find out more, join my Triple T session “DC Network Automation using K8’s constructs” on Wednesday, October 23 at 4:30 PM EST. Alternatively, stop by the Nokia table (#521) and we can show you more examples of how we can simplify network operations so you can confidently use automation to move at the speed of business.