A young engineer, Colin Cantrell, was on stage outlining the next evolution of blockchain technology at the 2017 Nexus Conference in Aspen, Colorado. As he stood on the stage, one of the biggest challenges facing the industry was scaling for mass adoption. Colin was presenting a new multi-dimensional chain that would make this a possibility, but there was still a question of how exactly nodes could scale on the network level. An audience member, Dino Farinacci, raised his hand, “Have you considered using IP multicast?” Colin had not, and they decided to meet after the conference to discuss this scaling solution that would not only greatly enhance the architecture, but also create a much needed level of stability at the foundational network level. That’s how the partnership between Nexus and LISP was born.
Dino Farinacci is a network engineer, the founder of lispers.net, and one of the original creators of the internet who worked as a Fellow at cisco Systems. He met Jon Hazell, a Nexus advocate and filmmaker, on a train, after attending a Bitcoin conference in 2014. Three years later, Dino was invited to the first Nexus Conference to present the history of networking. The event brought together some of the greatest thinkers and innovators in the world and has spurred a new era in the evolution of Nexus technology. As serendipitous connections go, this one led to a powerful breakthrough for both projects.
Dino teaches the Nexus team about LISP
To understand the benefits of this scaling network solution, we have to look at the difference between how the internet runs now, and what modifications are necessary to enhance its structure for newer applications. Most people assume the Internet already scales, but neglect to realize that when high capacity applications are in use, you load up the lower layers of the Internet infrastructure, which slows down message routing by using more resources. Did you know at the current rate of adoption of devices (cell phones, wireless earbuds, ipads, etc.) we have already run out of usable IP addresses?
IPv4 vs IPv6
The current internet infrastructure runs on the IPv4 system which can handle 4.3 billion IP addresses, while there are 8 billion possible users that own multiple devices. Designed in the early 1990s, IPv6 has 4 times the address space of IPv4, where every particle on earth could be addresseable. After all this time, the Internet is still transitioning from IPv4 to a dual-stack system, to ultimately IPv6. In the meantime, LISP can be used to get end-systems IPv6 addressed by using its overlay technology. This will allow more devices to be connected even when the core network underlay is not IPv6 deployed.
Unicast vs. Multicast
In Unicast flood networks all nodes must process and then relay information individually. Think of how you communicate with your friends about a concert you’re all going to. You can text the event information to a friend, and he texts the next friend, and he texts his girlfriend, etc., until everyone has the correct information. This uses more resources, and slows down information flow, but is extremely useful in cases like streaming movies where each user needs to have individual control over when to stop or pause. Multicast is akin to group text messages where you can text all of your friends at the same time. This increases speed and creates parallelism as the message is broadcast to all the relevant nodes at the same time, while preserving global consensus.
LISP is the Locator/ID Separation Protocol that separates IP address semantics into a fixed identity address (EID: End-point ID) and a locator address (RLOCs: Routing Locators). Instead of your identity changing each time you move your device, it stays fixed and obfuscated, allowing everything to run more smoothly. Comparing this to blockchain, your EID basically becomes the “trust key” for your identity. Your EID address is the same address you would be using today on your device, but the address can remain assigned to you independent of how and where you connect to the network. Instead of running on the IPv4 underlay, LISP send packets through on the IPv6 overlay, which adds another level of cryptographic security to the information. Most importantly, it creates a distributed topology regardless of connection method or location.
From Chance Meeting to Today
Colin and Dino have spent the first part of 2018 integrating the LISP architecture into the Nexus network and testing the Nexus nodes on the overlay. Nexus is currently developing its Tritium Software Stack: the first on-chain scalability upgrade to the Nexus blockchain. While scaling on the ledger level is crucial, network scalability plays just as an of important role. With the two protocols working seamlessly together, mass adoption becomes an attainable goal. This is the type of breakthrough that can only happen when people come together and collaborate to build a better future together.
Dino and Colin look forward to introducing the world to the new developments between network and ledger scalability, and the possibilities that arise when Nexus and LISP combine forces at the 2018 Nexus Conference. Please join us as we build the future together!