Uploaded November 2025 | Updated November 2026, 4 days ago
We share the story behind LinkedIn's Software Defined Backbone, one of the largest in-house networking systems built from the ground up. Faced with the demands of massive bandwidth across global data centers, we reimagined how modern networks should be designed and operated, not as monolithic systems burdened by generality, but as distributed systems purpose-built for flexibility, scalability, and control.
Our backbone leverages a SDN architecture, with centralized traffic engineering dynamically programming flows on cost-effective routers. This system drives link utilization to over 95%, balances traffic flows of different priorities across multiple paths, and adapts to real-time demand and failure conditions, all while delivering up to 5 Tbps of production traffic across sites.
We'll cover architecture, design, and reflect on three years of experience from deploying and operating this system in production at scale.
Jaffar Abdul: Jaffar Abdul is currently working as a senior staff software engineer at LinkedIn, where he has led many impactful projects that automate network provisioning, SDN based traffic engineering and host networking. He has over 17+ years of experience building networking products and deep expertise from networking operating system to network management software at scale.
Prashanth Kumar: Prashanth Kumar is currently working as Sr Staff Network reliability Engineer at Linkedin with 20+ years in the Industry. He has experience in building and operating large scale Backbone, DC and Edge Networking. Expertise spans network design, SDN/Automation, optical technologies, and disaggregated Network Operating System.
We share the story behind LinkedIn's Software Defined Backbone, one of the largest in-house networking systems built from the ground up. Faced with the demands of massive bandwidth across global data centers, we reimagined how modern networks should be designed and operated, not as monolithic systems burdened by generality, but as distributed systems purpose-built for flexibility, scalability, and control.
Our backbone leverages a SDN architecture, with centralized traffic engineering dynamically programming flows on cost-effective routers. This system drives link utilization to over 95%, balances traffic flows of different priorities across multiple paths, and adapts to real-time demand and failure conditions, all while delivering up to 5 Tbps of production traffic across sites.
We'll cover architecture, design, and reflect on three years of experience from deploying and operating this system in production at scale.
Jaffar Abdul: Jaffar Abdul is currently working as a senior staff software engineer at LinkedIn, where he has led many impactful projects that automate network provisioning, SDN based traffic engineering and host networking. He has over 17+ years of experience building networking products and deep expertise from networking operating system to network management software at scale.
Prashanth Kumar: Prashanth Kumar is currently working as Sr Staff Network reliability Engineer at Linkedin with 20+ years in the Industry. He has experience in building and operating large scale Backbone, DC and Edge Networking. Expertise spans network design, SDN/Automation, optical technologies, and disaggregated Network Operating System.










