Should you automate your optical network?
The question about the value of automation is often raised, especially now that it has become almost ubiquitous to talk about automation at every level of the network.
Is there value in automation? Yes. However, a short answer tends to oversimplify the topic, which is multi-faceted and complex because it involves many aspects of management and administration of optical networks.
One reason to automate is because automation reduces human error. But how do you account for the need to have oversight of the automation process? To be realistic, carriers are not likely to let their networks self-optimize any time soon.
Another value of automation is that it allows you to keep up with the increasing complexity of your network. However, automation can also impose new process complexity on the operations team. This new complexity may also require properly training the workforce to manage it.
Some people argue that automation will automate only repetitive tasks. Although this is a valid point, each customer is different and the repetitive tasks they leverage are also unique, so some customization is required. Common areas of interest with challenges include:
- Bandwidth-on-demand provisioning and service activation
- Network inventory
- Network configuration
- Predictive network health analysis.
Although there are ways to automate these tasks, each has its own niche challenges which may require some customization to tailor it to integrate into the IT ecosystem.
Fortunately, the optical layer compared to other technologies used in building networks is probably the one with the least amount of software hooks required as a baseline for automation. But the optical layer also probably has the largest number of physical layer proprietary features. This just adds to the complexity in automating optical networks.
So, there is obvious value in automation but also many challenges. Where to go from here? Let’s focus on the foundational elements needed in the optical network to enable automation—and address specific business pain points.
Zooming in on the automation elements
At Nokia’s Network Infrastructure division, we have been working diligently to simplify automation. Our approach relies on the foundational elements shown in Figure 1.
Figure 1. Foundational pillars for intent-driven automation success
Targeting business outcomes
The Nokia automation paradigm is based on an operator’s unique business objectives. These target business outcomes—focused solutions to meet an operator’s specific needs.
For example, one operator wants to run its network hotter, taking advantage of best performance on Day 1. The operator also wants to provide new, revenue-generating services with a quicker turnaround and to a multitude of business partners and subscribers. And the operator wants full visibility into the network to better forecast how it needs to scale to support the business.
To respond to the various needs of different operators, our automation solution features eight use cases that address automation of all aspects of both the network and services life cycle management. It’s important to remember that the Nokia solution always targets the business outcomes.
Our approach to automation is structured using the following building blocks to realize use cases:
This approach provides maximum flexibility to address the various use cases as well as the required level of customization to tailor the use cases for specific customers and specific business outcomes.
Using an approach based on business outcomes recognizes that there is no one-size-fits-all solution. As a result, the capability to customize solutions is a very important element of automation.
This flexibility provides an operator with options: customization of the feature sets, how these are put together, and customization of the delivery. These options address both the do-it-yourself (DIY) operator and the operator who needs full turnkey support along with customization—because flexibility matters.
Investing in network programmability
Efficient control of all the elements of the automation process is critical and is provided through” network programmability.” This essential component allows you, the operator, to control all aspects and phases of the automation process to first ensure it is adopted, and then allow for successful operation.
Our team of experts are continuously developing both core building blocks and workflows that can be further combined into more complex workflows to help customize a solution. These complex workflows can be re-used across many of the eight use cases to enable a business outcome.
By creating such workflows, operators essentially control all the aspects of the execution of the automation. Network programmability is what truly makes automation realizable.
Using mature, open API interfaces
Open initiatives are proliferating across the industry for good reason: they are a true enabler for automation. A fair number of open interfaces are usually coupled with a data modeling language such as YANG and offer the capability of various levels of abstraction.
While YANG data models make machine-to-machine communication more efficient, abstractions are the true engine that allows operations support system (OSS)-level automation. This is because it is easier to configure a model than a physical device, and the higher in the network layer stack you go, the more abstraction will be required.
This abstraction usually translates into terms like model-driven interfaces and intent-based networking, which allows both configuration and operation with a minimum set of instructions. You can trigger an outcome without needing to trigger all aspects that go into making that outcome.
At Nokia, we have been investing heavily in open interfaces. We were the leader in the development of the ONF TAPI northbound interface. We were also extensively involved in the development of NETCONF and OpenConfig YANG data models for the southbound interface.
Evolving hardware capabilities through technology innovation
Our long experience in introducing new technologies within our hardware development cycle is key to the success of our automation approach. Led by Nokia Bell Labs, our teams were early adopters of streaming telemetry—the continuous and fast collection of network key performance indicators—from network equipment. Our teams have also been shaping our hardware using OpenConfig data models and NETCONF/gNMI interfaces.
Leveraging both technologies along with fast measurement of optical parameters is key to enabling analytics in the optical layer. With our innovative Nokia Bell Labs paradigms for smart sensing, this helps us build a foundation to automate large aspects of the network operation.
The bottom line
By using the foundational pillars of Nokia intent-driven automation, you can accelerate business revenues and lower the total cost of ownership of your optical network. Future blog posts will discuss specific use cases.