Take charge of your data center switching infrastructure
In the current era of cloud and as-a-service, there has been a shift in data center applications from classic client-server delivery models to microservices based models. Instead of the emphasis being on north-south traffic patterns, east-west flows are becoming critical, which favors leaf-spine network architectures and Layer 3 IP fabric designs. This shift also favors solutions using merchant-silicon based hardware and a new approach to network operating system (NOS) design. This shift is opening up opportunities as data center operators re-think the role of the data center network and how to have more freedom to innovate. This is putting disruptive players like Nokia in a thought leadership position as an integral partner in data center fabric enhancement and modernization projects.
The key advantage offered by the new approach is, of course, scalability. This is especially critical as the entire service chain is being driven from the top down with cloud applications calling the shots. Previously the network was a limited resource that applications had to accommodate. Today networks are much more flexible, both within the data center and in the wide area, as a result of virtualization and cloudification. Networks not only need to scale on demand but also need to adapt to evolving application needs in an agile manner. This requires operations-focused job functions and tasks to be performed in an automated way wherever possible. This approach offers service assurance that can be monitored in real-time, and by using AI/ML-driven approaches, they can also repair faults or route around them before the service is interrupted.
Instead of being a fixed variable in the service chain, networks are now a dynamic part of end-to-end service delivery. Service provider and enterprise operations teams can adapt enhanced NetOps approaches by leveraging DevOps in their innovation and customization pipeline — not only continuously integrating and delivering new application fixes, features and releases, but also continuously testing the network’s ability to deliver them. Automation can now be extended to all phases of the operations cycle including design, deployment and day-to-day operations, to ensure the highest level of freedom and flexibility for data center operations teams.
When we first decided to enter the data center switching market, our research told us over and over that data center operators were frustrated by how much the incumbent systems handcuffed them to a certain way of doing things, often feeling locked-in to vendor implementations or roadmaps. Increasingly, they wanted to break out and have the freedom to take charge of their operations. What their frustration reflected was the fact that their customers wanted cloud-like service delivery and response time and it was imperative for the data center network to follow the same approach and be cloud-like in the true sense.
This realization gave us our mission and focus in developing our data center switching solution. We committed to designing a solution from the ground-up that rearchitected the “status quo” to create platforms that were open, extensible, reliable, and highly customizable.
Our first step was to develop the most advanced NOS on the market, built on a Linux foundation and designed for openness and unmatched extensibility — a NOS for the cloud-native era and beyond. We adopted the latest in management protocols and a completely modular, model-driven approach. Each network application within the NOS operates independently and has its own YANG data structure. This foundation allows automation tools to deliver consistent automation across all systems, as well as streaming and on-change telemetry via gNMI for delivering unseen levels of granularity.
We leveraged our 20+ years of experience in IP protocol support by baking in protocols (e.g., BGP, EVPN, VXLAN) from our proven SR OS, so our customers have data center switching platforms that are robust and interoperable out of the gate, alleviating the concerns of bringing in a new vendor.
Our fabric management and automation platform is built on Kubernetes, which provides an extensible, containerized and microservices-based system with full access to existing Kubernetes utilities and applications. The system includes an integrated “network digital twin” capability that emulates the production network for trialing and validating any changes planned for deployment in the production network, speeding up testing and validation and reducing risk.
This platform supports intent-based automation for several types of intent-based templates, including design intent, workload intent and maintenance intent. This enables operators to define the desired end state of the fabric across Day 0, Day 1, and Day 2+ in an abstract manner. The system is easy to use as all features are connected with intuitive workflows that allow staff to ramp very quickly to embrace advanced and powerful automation capabilities.
Our data center fabric solution is helping our customers to take charge of their data center operations in enterprises, communications services providers (e.g., telco cloud buildouts), cloud service providers, colocation providers, hyperscalers and managed services providers.
Nokia’s innovative leadership was recently recognized in a GigaOm Radar Report on the data switching market. As GigaOm stated “Nokia 7200 IXR series switches deliver high-performance hardware with a fully programmable NOS built for NetOps from the ground up. The vendor ranks high on a variety of metrics, including automation and support for network design, deployment, and operations.”
This is the kind of innovation platform our customers have been asking for, and their response has been overwhelmingly positive, praising it for the freedom it gives them to provide added value to their customers. We are happy to deliver a true next generation solution for the data center switching market.
Please check out the GigaOm Data Center Switching Radar Report to see how Nokia is an innovation leader and learn how you too can take charge of your data center switching fabrics.