When you think about the next 5-10 years of Internet development, how do you see the network architecture changing?
Primarily we are going to see changes that move us away from the “best-effort” legacy of the last 20 years and towards a future that implements and places a greater value on a highly available architecture and the infrastructure that supports it. We are already seeing this happen and we will continue to see a push for this. Of course, this requires support from so many different components (e.g., fiber routing, power grid, protocol design, etc.) but the current momentum will continue to bring us all down this road.
Do you see the need for additional points of peering and network interconnectivity to be established beyond what is in place today?
That is an interesting question. My first instinct is to say yes – and I believe it to be the case in a number of places in the world today – but I believe the answer has moved closer to “that depends” in the last 5-10 years. The world’s current peering infrastructure is rather robust. While there are obviously still sizeable parts of the world where connectivity is sub-optimal, a large percentage of the global population is within a sub-100ms envelope to their content and data. The need for additional connectivity will be driven through a combination of technology improvements (e.g., latency improvements for mobile users from 5G) and evolving business models between platform owners, such as access and content platforms, and I think this is one of the items the new Richmond NAP is trying to address.
You have been instrumental in bringing the new transatlantic cable MAREA to Virginia which is now accessible from the NAP. What is the significance of diversity in subsea landings?
This goes back to the first question — diversity is critical to improving the overall availability of the global Internet infrastructure. But it also extends beyond a cable landing point; where network connectivity of any type ultimately terminates also plays a key role. Diverse landings that still concentrate in a single facility decrease the improvements such diverse systems create. Of course, diversifying the landing is great but you must follow the traffic flows and work to eliminate areas of concentration that increase the risk of outages and downtime. Infinite availability to products and services from anywhere in the world is what users have now come to expect and demand from the digital economy.