Recently, we hosted more than 500 engineers at the Computer History Museum in Mountain View, California for our fourth Networking @Scale conference. As with previous Networking @Scaleconferences, w…
Overview
The 2018 Networking @Scale conference brought together over 500 engineers to discuss advancements in networking technologies, featuring speakers from major companies like IBM, Microsoft, and Facebook. Key topics included automation in network provisioning, optical networking challenges, and the importance of IPv6 adoption.
What You'll Learn
1
How to automate complex network provisioning with systems like Vending Machine
2
Why machine learning is essential for modeling network design intent
3
How to implement a scalable Layer 4 load balancer using Katran
Key Questions Answered
What advancements were discussed in high-speed optical networking?
Alan Benner from IBM highlighted how supercomputing has addressed high-speed optical networking challenges, emphasizing the need for automation and tools to deploy and debug dense optical networks. This is crucial for building high-speed data center networks, where constraints like power and space are significant.
How is Facebook automating network provisioning?
Facebook's Vending Machine system automates complex provisioning processes, such as configuring new Edge Points-of-Presence. By streamlining these procedures, it ensures that all team members are updated on the necessary steps, significantly reducing manual errors and time.
What is the current status of IPv6 adoption according to Facebook?
Facebook reported that IPv6 traffic from the U.S. recently surpassed 50%, with mobile traffic exceeding 75%. This highlights the growing importance of IPv6 as more countries see rapid increases in adoption, driven by specific providers.
Key Statistics & Figures
IPv6 traffic from the U.S.
50 percent
This statistic indicates the growing reliance on IPv6 as part of Facebook's network infrastructure.
IPv6 mobile traffic
over 75 percent
This highlights the significant uptake of IPv6 in mobile networks, reflecting broader trends in internet connectivity.
Technologies & Tools
Backend
Katran
A scalable Layer 4 load balancer that Facebook has open-sourced to improve network performance.
Backend
Vending Machine
A system developed by Facebook to automate network provisioning processes.
Software
Machine Learning
Used by Microsoft to model network design intent and optimize resilience.
Key Actionable Insights
1Implementing automation in network provisioning can drastically reduce errors and improve efficiency.By using systems like Vending Machine, teams can streamline complex configurations, ensuring that all members are informed and reducing the risk of mistakes during network setup.
2Understanding the challenges of optical networking is essential for future-proofing data center designs.As data demands grow, addressing power and space constraints in optical networking will be critical for maintaining high-speed connectivity.
3Leveraging machine learning for network design can enhance resilience and minimize downtime.By running simulations to identify high-risk areas, teams can proactively optimize network links, ensuring robust performance even during failures.
Common Pitfalls
1
Failing to automate network provisioning can lead to significant errors and inefficiencies.
Without automation, teams may struggle to keep track of complex configurations, leading to mistakes that can disrupt network operations.
Related Concepts
Networking Automation
Optical Networking Challenges
Ipv6 Adoption Trends
Machine Learning In Network Design