In a move to address the growing complexity of cloud networking, the Linux Foundation launched Paraglider, an open source initiative designed to simplify setting up and managing networks across multiple cloud providers.
Cloud networks have long been plagued by complexity, often requiring users to navigate intricate low-level networking components to achieve their desired connectivity goals. Paraglider aims to manage this complexity by exposing a high-level API that aligns directly with the tenant's connectivity intent, allowing users to focus on what their networks are trying to achieve rather than the intricacies of network assembly.
The Paraglider project is a collaborative effort involving major industry players such as Aarna Networks, Broadcom, Google Cloud, IBM, Intel, Microsoft, Uber and UC Berkeley. By bringing together expertise from these diverse organizations, the Linux Foundation aims to redefine how cloud customers provision and manage their networks in multicloud environments.
“Paraglider simplifies the interface to cloud networking resources,” Sarah McClure from UC Berkeley and part of the Paraglider team told SDxCentral.
Paraglider aims to 'glide' past modern cloud networking complexityParaglider is looking at the issue of cloud networking from a different perspective than other efforts.
McClue said that some existing options for setting up cloud networks are using the underlying cloud APIs directly or using something like Terraform. While Terraform has its benefits in being infrastructure as code, in her view it doesn’t solve the underlying problem. That problem is that organizations are still speaking in terms of low-level components like virtual networks, gateways, etc.
She added that other options in the space like service meshes can simplify connectivity between apps at the application layer, not the network layer. Often approaches assume that someone else has set up the IP-level connectivity and then uses that underlying connectivity to create application-layer constructs like service meshes.
The promise of Paraglider is its capability to provide a unified, simplified API across multiple clouds, enabling users to directly specify their network-level connectivity needs. By seamlessly creating the necessary networking infrastructure based on the tenant's high-level intent, Paraglider eliminates the need for users to worry about the technical details of network configuration and inter-cloud connectivity.
How Paraglider will work with Kubernetes and other cloud-native technologiesThe Linux Foundation is already home to a variety of cloud technologies including Kubernetes via the Cloud Native Computing Foundation (CNCF).
McClure said that Paraglider is planning to work towards integrating with kubernetes more into the future. However she noted that Paraglider can support many cloud resources that are not just clusters. McClure said that many of the CNCF current cloud connectivity efforts assume there is underlying IP-level connectivity.
“In fact, short of making all nodes public, there is not much kubernetes can do to help with setting up network-layer connectivity,” she said. “For example, if you need to connect two clusters in two different private virtual networks, you will need something to set up the connectivity between those networks.”
The Linux Foundation is also home to the LF Networking group which is an umbrella organization of multiple open source networking projects.
“We are not yet a part of LF Networking, but joining is a next step,” McClure said.
McClure said that other efforts make different assumptions about the setting and workloads for networking connectivity.
“I think Paraglider could be leveraged by these projects to achieve network-layer connectivity more easily,” she said. “We're excited to work with the community to find points of collaboration.”
Comments