What Network Operators Can Learn from How Ford Beat Ferrari

This Industry Viewpoint was authored by Wayne Hickey, Advisor, Product Marketing, Ciena

A blockbuster production of “Ford v Ferrari” starring Matt Damon and Christian Bale recently hit theaters, raking in nearly $185 million thus far and soon contending for Oscars in the new year. More than just popcorn fare or awards fodder, though, it can serve as a lesson for network operators, too.

* Some spoilers ahead (if a true story from 1966 can be spoiled) *

As his race team embarked on the first of a string of victories (1960-65) at the famed 24 Hours of Le Mans, Enzo Ferrari said, “aerodynamics are for people who can’t build engines.” In 1966, Carroll Shelby designed a car for Ford that would end Ferrari’s winning streak. Rather than just building a bigger engine, however, Shelby used technicians from a Ford-owned aerospace company to rig a computer into the passenger seat and used Scotch tape to attach yarn to the outside of the driver’s side of the car.

Shelby, his team, and Ford used old- and new-fashioned data analysis to find that they were losing horsepower due to poor air ducting and spent weeks redesigning that system – not the engine – to provide maximum horsepower and torque.

Network operators have historically used the Ferrari model. They’ve relied on bigger routing fabric, more memory, and more compute power to route more and more packets to IP-enabled destinations. But that’s about to change. Operators are simplifying the delivery of end-to-end services across segmented network infrastructure to address unprecedented demand in services and scale. They’re turning to the Shelby/Ford model by leveraging dynamic routing technology.

Moving packets from source to destination across multiple networks is complex and difficult. Adding more and bigger routing “engines” doesn’t improve delivery efficiency. If anything, it makes matters worse.

Traditional IP Routing

The challenge inherent to traditional IP routing is the lack of defined structure. If you were to track IP packets as they move within and across traditional IP networks, you’d be hard-pressed to find any logical systems or patterns – because there aren’t any. When a routing engine (CPU) receives an IP packet, it doesn’t come with any instructions or path, just the final destination. This means that each CPU is making independent forwarding decisions for each and every packet, which is repeated at each router hop until the packet eventually reaches its destination address.

Clearly, this isn’t a picture of efficiency. All of these packet hops, and individual routing decisions manifest in buffering and poor, unpredictable performance. This simply isn’t viable for time-sensitive applications or next-generation network applications like UHD video streaming, 5G, or IoT. Putting a “bigger engine” –more processing power – into this process will not solve this inefficiency problem; network modernization will.

Static vs. Dynamic: Advantages & Disadvantages

Static routing allows network operators to tell every static router the next hop received IP packets must follow, which means manually provisioning “all” known routes. It doesn’t require high levels of processing power or overhead, and works for small networks, but the operational complexity and expense of scaling a network based on static routing make it impossible from cost and service velocity perspectives.

Instead, operators are turning to dynamic routing to achieve the operational simplicity required to scale reliably, efficiently, and at the lowest cost. Here, they configure routers with protocols that automatically learn available routes by communicating with and learning from other routers to understand the holistic network topology. This dynamic routing approach is broken into two separate routing protocols: Interior Gateway Protocol (IGP) and Exterior Gateway Protocol (EGP).

While IGP protocols route traffic within an Autonomous System (AS), which is a distinct IP network, EGP routes traffic between these ASs. These routing protocols require more processing power and have higher overhead than static routing. However, the operational simplicity, change notifications, and aforementioned automated routing updates make it the only choice for networks to properly scale.

There is a scaling challenge when it comes to IGP and very large networks delivering End-to-End (E2E) services via dynamic routing. IGP systems aren’t designed to support hundreds of thousands of IP prefixes, making it expensive and complex to manage, maintain, and troubleshoot IGP systems stitched between IP domains.

Here we turn to seamless and Segment Routing (SR) Multiprotocol Label Switching (MPLS), which groups all services together as an E2E Label Switched Path (LSP), or segments. This simplifies routing by leveraging EGP, which is designed to support the Internet and its untold number of routes.

Seamless MPLS, Segment Routing & An Aerodynamic Future

While operators traditionally build a wireline network and a wireless network, each with multiple Ethernet and MPLS aggregation architectures, converging access, aggregation, metro and core to a seamless MPLS network simplifies provisioning, operation and maintenance. This underlay eliminates manual stitching, ensures seamless, boundary-less connectivity and, as a result, offers true E2E connectivity. Ethernet, MPLS, Level-2, Level-3 and private line services no longer need to be separated. Rather, they’re connected directly to the core.

Another movement for IP network simplification is Segment Routing, which chooses a path and embeds it directly into the packet header as an ordered list of links. This eliminates the need for nodal signaling and protocols making the network much simpler to own, operate, and troubleshoot. These efficiency- and simplification-driven approaches to the modern network are in the same spirit as the aerodynamic efforts the Carroll Shelby made more than a half-century ago – though not the only factor in winning at Le Mans or creating a dynamic IP network.

History has shown that horsepower and aerodynamics play vital roles in each endeavor. While horsepower enables speed, aerodynamics makes the car more drag efficient and provides down force. Down force improves the car’s grip on the road, but also increases drag, thereby limiting top speed. One way of reducing drag is to apply active aerodynamics, like an openable rear wing to reduce drag and downforce when you don’t need it.

Ultimately, a high level of performance is neither a function of having the most power nor the highest level of complexity. Ford’s 1966 champion GT40 was powered by a seven-liter V8 engine and made 360 laps at The 24 Hours Le Mans race; 2019’s champion completed 385 laps with a 2.4-liter V6 engine. The best of years ago may have been good enough then, but those approaches cannot stand up to the latest technology, whether in racing or networks.

Some vendors may want to lock network operators into the routing protocols of yesteryear, but as innovation continues, those operators may be left in the rearview mirror of their competitors. The ability to aerodynamically route existing and new services using seamless MPLS and SR-MPLS at scale doesn’t just result in faster speeds but reduces network complexity and provides higher levels of service velocity.

Traffic forwarding is changing, and networks are ready to reduce drag and connect all services across a single network, providing a migration to the future for legacy services, and enabling new opportunities. Carroll Shelby’s Shelby American race team did it for Ford more than 50 years ago.

Operators have the ability to do the same for their networks.

If you haven't already, please take our Reader Survey! Just 3 questions to help us better understand who is reading Telecom Ramblings so we can serve you better!

Categories: Industry Viewpoint · Telecom Equipment