Building Waze for The Next Internet

By Special Guest
Jonas Simanavicius, CTO, NOIA Network
  |  July 22, 2020

Nobody likes traffic. Sitting in your car on the highway after a long day at work is about as dull of an activity as there exists. It’s frustrating, and before GPS-linked applications emerged, there was no way of looking at real-time traffic other than what was briefly detailed on the radio. Eventually, Google (News - Alert) Maps and other applications came to the rescue, providing a birds-eye view of your planned route. 

However, traffic is not always avoidable, nor do GPS apps always divert your path away from slowdowns. That’s where Waze built a cult-like following among its early users, eventually snowballing into 115 million users with user-driven alerts and traffic updates. 

In particular, Waze navigates users through traffic bottlenecks with obscure routes – sometimes through a maze of back roads – all in the name of avoiding bumper-to-bumper traffic. Ironically, even though Waze relies on GPS and a mobile app, we have yet to deploy the same concept to an increasingly precious landscape of digital streets – the Internet. 

Internet speeds from ISPs are slower than promised, enterprise outages cost more than $700 billion per year, and congestion on the Internet is endemic. Throw in millions of people suddenly working from home, and it begs the question, Why haven’t we designed a Waze for the Internet? 

Internet Protocol & Routing – Bypassing Congestion 

The Internet is much older than it appears behind the sleek social media applications and enterprise dashboards most users see today. Conceived in the 1970s, many of the Internet’s critical, underlying protocols were created decades ago and still form lynchpins of the Internet’s highways today. 

For example, the Border Gateway (News - Alert) Protocol (BGP), the primary path-vector routing protocol on the Internet, was launched in 1994. BGP is designed for ISP routers to discover pathways for Internet traffic to various IP addresses, whether they are users at home watching Netflix or enterprise customers plugging into a cloud service. 

However, BGP can only direct traffic to the next destination or neighboring router and is incapable of optimizing the entire route. Like conventional car traffic before GPS technology, this would be akin to a car driver following a physical map and relying on sparse radio updates on what the traffic looks like. Alternative routes are not provided, and a real-time display of the traffic overview on their path is not available. 

BGP directs significant chunks of Internet traffic via the same routes, creating significant points of congestion and connectivity failure. 

One of the immediate problems that arises from BGP’s limitations is that technology companies, such as cloud service providers, cannot guarantee connectivity reliability due to the localized, fluctuating nature of BGP. Additionally, innovating around the problem is cumbersome without relying on layered technology, like overlay networks, to improve connectivity. 

SaaS (News - Alert) providers subsequently have to charge higher fees for their services to cover connectivity issues and the associated maintenance. 

With everyone suddenly working from home in March, questions about the Internet’s ability to handle traffic surges naturally arose. The response from major ISPs was, “yes,” they could manage the spike in broadband activity for homes, but the abrupt shift in Internet behavior raises some forward-thinking questions at the enterprise level. 

Namely, is the current Internet infrastructure for enterprise-grade activity sufficient to support the onset of the edge computing revolution, which is closely tied to the IoT market. BGP’s current iteration would indicate no.  It’s time for the Waze of Internet routing to step up. 

Waze for the World of IoT and Edge Computing 

IoT technology requires optimized routing at the network “edge,” meaning that edge data centers, which are expected to account for a $16 billion market by 2025, will progressively increase in demand alongside the growth of smartphones, IoT devices (consumer and enterprise), and cloud-based services. 

Edge computing brings storage and computation closer to the “edge” of the networks, such as drone-based farming networks in rural areas. It  is closely linked to the rise of IoT devices, whose sensory component in devices ranging from electric cars to real-time shipping tracking comprises a significant portion of SaaS enterprise activity.  

Current data center architecture simply won’t cut it for the surfeit of smart devices that are set to contribute to a ballooning digital economy. The demand for edge data centers is primarily driven by a new generation of OTT services, where high bandwidth and low latency are the primary appeals – avoid congestion issues of traditional high-density data centers towards the edge. 

But enterprises switching to edge data centers is costly, not only from the expense of infrastructure development, but maintenance with some of the same problems as before. 

Tinkering with new routing optimization requires a more bottom-up approach, tackling the plumbing of the Internet’s highways head-on. That’s where replacing BGP can make an enormous impact. Fortunately, significant research and development with the Internet’s most recent protocol version (IPv6) have opened up a distinct opportunity for the Internet’s Waze.  

Combining IPv6 with Segment Routing, a method for routers to understand packet information and destination, a new routing protocol, SRv6, can actualize. With SRv6, the neighbor/local forwarding obstacles of BGP are removed, as networks can be programmed at the packet level and drive the concept of crafting a Waze for the Internet.  

From a broader perspective, this means that Internet traffic can be routed around bottlenecks, with alternative avenues programmable by enterprises. Maintenance and infrastructure costs are vastly reduced. When combined with a distributed ledger (i.e., a blockchain), SRv6 even enables public Internet infrastructure to perform like private networks by decentralized indexing of IP locations.

Rather than blindly pushing high-density and edge computing Internet traffic through congested highways, traffic is efficiently routed around the hubs of obstruction. Latency improves, as the new protocol can continuously measure the performance between the primary route and alternative routes, taking account of which path is fastest. 

Considering the vast growth of IoT traffic expected in the coming decade, a hybrid SRv6 and distributed ledger platform can serve as the Waze of the Internet’s next iteration. 

The costs of not revamping the underlying protocols of the Internet are simply too high to ignore before unleashing the next wave of IoT technology. 

Imagine returning to the days of physical maps and relying on car radio traffic updates. It probably makes you cringe. Waze has made daily commutes and road trips a little less onerous now it’s time we do the same with Internet traffic. The next iteration of the digital economy depends on it. 




Edited by Erik Linask