Posted on Reading Time: 4 minuteson
Software-Defined Wide Area Networking (SD-WAN) popularity in the networking world has been constantly on the rise, mostly thanks to the unprecedented user- and application-directed control offered over network resources and service capabilities. Meanwhile, in the application world, it is almost impossible to have a conversation without mentioning Kubernetes, given the flexibility and robustness it brings when hosting applications at scale.
How do these two rising technologies, SD-WAN and Kubernetes, relate to each other? What benefits arise from deeper integration between them? In this article, we delve into those questions with the perspective brought by the MEF 3.0 SD-WAN standardization initiatives.
SD-WAN for Kubernetes
SD-WAN is becoming a common way for enterprises to handle their WAN connections, given the advantages of flexibility and cost-effectiveness that come from adopting a software-defined network architecture. SD-WAN policies allow network operators to fine-tune a WAN overlay to make the most of a set of underlay networks.
On the other hand, Kubernetes is the de facto standard for newly developed applications, providing an extensible and robust way to efficiently host applications at scale. However, modern distributed enterprises often find that the Kubernetes clusters hosting their applications are not collocated with the users or applications that consume them. As a consequence, it is not uncommon to find that the connectivity to the Kubernetes clusters is provided by some form of SD-WAN solution. While these two technologies are often deployed together, they tend to be unaware of each other, missing out on optimization opportunities.
Cloud Native SD-WAN
The Cloud Native SD-WAN (CN-WAN) project is an open-source initiative that integrates SD-WAN with Kubernetes, recently announced at KubeCon, the premier Kubernetes conference. The goal of CN-WAN is to make SD-WAN services easily consumed by the applications hosted in a Kubernetes cluster. In a nutshell, CN-WAN takes the application network requirements, as expressed by DevOps as part of the Kubernetes application specification, and maps them to SD-WAN policies, as defined by NetOps using the SD-WAN orchestration system. In this way, CN-WAN provides to the NetOps the tools to automatically optimize the network resources needed by the applications deployed by the DevOps.
NetOps for DevOps
One might wonder, what are the advantages introduced by CN-WAN, and why is such integration of SD-WAN and Kubernetes integration needed? In many deployments today, the application infrastructure uses the network as a pure transport without leveraging its full potential. In some other deployments, the underlying network optimizations fall short due to the lack of visibility into application requirements. In this context, the CN-WAN project builds on top of two simple observations:
- Optimal application experience is only possible with proper network optimization.
- Applications and networks are two dissimilar knowledge domains (DevOps vs NetOps).
The CN-WAN initiative is developing on the basis that:
- Network operations play a critical role in delivering optimal application experience.
- Infrastructure and service providers are crucial to ensure that the network is optimized for each application.
The project also assumes that it is challenging for DevOps to know how to make the most out of the network, and similarly challenging for NetOps to fully understand the requirements of a given application. With that perspective, CN-WAN aims to provide a clear demarcation and interface between DevOps and NetOps (i.e., between Kubernetes and SD-WAN).
Following that clear demarcation, the service provider NetOps team, aware of the specific characteristics of the underlying infrastructure and network deployed, creates mappings of possible application metadata to different network policies. In this way, NetOps abstract the complexity of fine-tuning the network and just expose a set of easy-to-use application attributes. DevOps can then use these abstracted attributes when defining their applications.
As a result, CN-WAN effectively makes the SD-WAN NetOps optimization services easily consumed by DevOps, thereby automating the delivery of optimal application experience.
CN-WAN for MEF SD-WAN Services
The MEF 3.0 SD-WAN service standard, as set out in MEF 70.1 Draft Release 1 SD-WAN Service Attributes and Service Framework, provides a unique opportunity to streamline the consumption of SD-WAN services by Kubernetes applications.
- On the one hand, service providers can ensure a way for the advanced capabilities of their underlay networks to be consumed by the upper layers of the stack, thus differentiating their infrastructure offerings.
- Conversely, service providers can easily expose the features of their underlay services, as part of SD-WAN policies tailored for specific application requirements.
Extending MEF 3.0 SD-WAN service standards to address cloud native application workloads, as enabled by CN-WAN, is particularly important for those service providers that offer managed SD-WAN services. In fact, thanks to CN-WAN, those service providers can offer automated SD-WAN optimizations per Kubernetes application deployed by their customers. This ability results in a model for easily consumed SD-WAN managed services in which customers just express their application networking needs with simple Kubernetes configurations and the managed SD-WAN solution can then automatically and transparently render the appropriate network optimizations.
To see all this in action, please visit the upcoming MEF 3.0 PoC (136) MEF SD-WAN Service Provided Using Cloud Native SD-WAN showcase on 22 February 2021, which demonstrates a MEF-standard SD-WAN service using open source Cloud Native SD-WAN provided by Cisco, Equinix and ngena CN-WAN. The showcase will illustrate how, thanks to CN-WAN, Cisco Viptela-based SD-WAN services managed by ngena automatically provide per-application WAN optimization to a cloud-based Kubernetes cluster, connected via Equinix Cloud Exchange. Read more about MEF 3.0 PoC (136) >
Also see the Cisco blog article, Cloud-Native SD-WAN: The WAN Your Kubernetes Applications Deserve
About MEF’s Proof of Concept Program & Showcase
The MEF 3.0 Proof of Concept program effectively fosters innovation, seeds new MEF standards and projects, and accelerates our existing work within the ICT industry by providing our members—including service providers, technology suppliers, and other stakeholders within the ICT industry—the opportunity to collaborate on MEF 3.0-based use cases throughout the year.
Initiated by MEF members and facilitated by MEF staff, each MEF 3.0 PoC receives a unique, identifying number that remains unchanged as its title and messaging evolves over the life of the project.
PoC work is highlighted in public showcases and award presentations that explore individual Proofs of Concept.
Alberto is a Technical Lead at Cisco’s Intent-Based Networking CTO team, where he focuses on the intersection of Enterprise Networks and Cloud. Lately he has been leading the “Cloud Native SD-WAN” project, an open-source initiative to improve the integration of SD-WAN and Kubernetes. Before Cisco, Alberto earned a PhD in Software-Defined Networking at BarcelonaTech.
With contributing authors, Fabio Maino and Charles Eckel.