The Networking Nomad

There were some significant announcements made by Nvidia and Amazon Web Services (AWS) during the Nvidia GTC event in San Jose, showcasing a comprehensive collaboration to advance AI innovation and infrastructure globally. 1. Project Ceiba:  The most notable announcement is Project Ceiba, a joint initiative to build one of the world's fastest AI supercomputers hosted exclusively on AWS infrastructure. Utilizing Nvidia GB200 Grace Blackwell Superchips, this supercomputer aims to push the boundaries of AI research across various fields such as large language models, digital biology, autonomous vehicles, and climate prediction. 2. New Blackwell EC2 Instances:   AWS plans to integrate Nvidia's advanced Blackwell GPUs into its Elastic Compute Cloud offerings, introducing a new class of EC2 instances tailored for modern AI demands. These instances are designed to handle complex AI tasks, leveraging Nvidia's GPU technology with AWS's scalable cloud infrastructure. 3. AWS Nvidi...

In this lab, we will see route preference at play in Huawei routers. Consider this Topology: According to the routing preference guide: The higher the preference number the lower precedence given to the protocol. If Route is learned from iBGP, it will have the lowest preference out of the given options.  Let’s configure BGP on both sides and see. On R1 along with configuring BGP, advertise the Loopback Interface as well When the peer is established, we see that the route is being learned by iBGP and has the highest preference number. If we configure OSPF and redistribute the Loopback address on CX1 the route being learned from iBGP will be replaced by OSPF external route entry. Let’s configure OSPF on both sides and import the loopback interface. A prefix list is implemented so that only the desired route is redistributed.   Even though the BGP route is still valid, in the routin...