At Dell Technologies World last week in Las Vegas, the company made lots of announcements, which have been beautifully covered by the hundreds of analysts and journalists who attended. For my part, I would like to highlight the progress of two Dell partners whose work struck me personally: McLaren and Briggo. The former is known for its Formula 1 racing cars. The latter isn’t known yet — but will be — for its barista robot.
I’m not much of spectator-sports guy. I attended my first professional basketball game a few years ago, and I think I’m good with that. I’ve never been to a pro football or hockey game, and that’s not likely to change. I will admit to a slight weakness for baseball, which can perhaps be explained by the fact that I grew up in Red Sox territory. But at the level of individual sports — cycling, skiing, sailing, and others — I’m all in.
So, McLaren is not an obvious fit for me. However, stopping by their booth on the show floor, I met Paul Brimacombe, head of IT architecture, and while others ogled the car on display, we talked about driving algorithms. It turns out, Brimacombe has done many of the same thought exercises that I have about driving protocol in general and how self-driving vehicles should behave in particular. That conversation led to another over drinks at a partner-analyst cocktail party that evening. Turns out, McLaren works with the likes of Google’s DeepMind subsidiary and universities on artificial intelligence problems like self-driving.
Driving with Data
First, a bit about why McLaren was on the show floor in the first place. Formula 1 is a rarefied form racing involving specially built single-seater cars. These cars have more in common with rocket ships than that thing in your garage. And it turns out that success depends on a tremendous amount of data analysis. The cars themselves are heavily instrumented, and these sensors spin off huge buckets of data in real time. This data is used to improve car design as well as to manage the car during a race. Also, data drives how the drivers drive. The courses themselves are instrumented so that handlers can analyze the field of vehicles and advise their drivers on optimal strategies over a radio connection to an earpiece. McLaren uses Dell gear to do all this data crunching.
Across the McLaren Group, Dell EMC VxRack Flex systems and Dell EMC M1000e chassis and modular servers provide most general-purpose computing. The VxRack Flex systems deliver block storage via storage-area networks (SANs), Dell EMC Isilon handles network-attached file storage (NAS), and Amazon Elastic Container Service (ECS) is used for objects.
At trackside, the company predominantly deploys Dell EMC PowerEdge R740xd servers to get a blend of processor and disk performance in a hyperconverged package.
Most engineers and analysts work on well spec’d Dell Precision 5530 and 7730 mobile workstations, usually with Intel i7 processors and a minimum of 32GB of memory; other staff do their jobs on Dell’s Latitude family of notebooks, with the 7000 series proving most popular, according to Brimacombe.
Here’s the bit that ties into my interests: Brimacombe told me over beers that McLaren has just bought a half interest in the Bahrain Merida Pro Cycling Team. Now, pro cycling is a team sport, but I actually covered the Tour de France for Bicycle Guide magazine back in the mid-1980s when I was cyclo-touring through Europe. With my long lens, I caught Bernard Hinot and Greg Lemond dueling it out on the second-to-last hairpin of the L’Alpe d’Huez climb in 1986.
I perked up when Brimacomb mentioned McLaren’s new interest. Turns out, cycle teams operate a lot like F1 teams. There’s a designated leader and a pack that tries to protect him. Many of the same strategies are deployed: block a challenging rider, speed up to try to lose him, or call in a domestique to attack the challenger. These decisions must be made rapidly. Is the challenger in the top five or merely an upstart from down the ranks who can be ignored? Does the leader need to conserve energy for a climb ahead, or can he respond to the challenge and still have enough in the tank to win on the climb? All these decisions are improved with data properly crunched. Bahrain Merida will help promote McLaren’s brand beyond its traditional fan base.
But cycle racing has some interesting data problems that F1 doesn’t. On an F1 course, the entire circuit can be instrumented and networked. A car is never out of range of a transceiver. But in the mountains of Europe, cellular coverage is sketchy at times, and WiFi doesn’t have the range. Thus, McLaren is grappling with how to connect its cyclists to the same type of data machine its drivers enjoy. Brimacomb highlighted the choices and issues. First off, the team can be networked via mesh links as long as the members are not too far apart. But at least one rider has to be connected to a wide-area router to transmit signals back and forth to and from the data store. Of course, some things can be done locally (i.e., right on the bike) but many require a trip to the cloud. The type of data we’re talking about here are rider performance stats like heart rate, breathing, and pedal cadence; mechanical information like gear, lean angle (on turns), and tire pressure; and course data like time elapsed, rider position, and approaching topography. The crunching could be, knowing this rider’s historical stats, does he have enough reserves left to attack on the next climb?
A helicopter could follow the team, but the whop-whop-whop of the chopper blades could be distracting for riders and unpleasant for spectators. A balloon or blimp represents another possibility, but control and reliability could be problematic. Brimacomb seemed to be leaning toward a following car, which could network with the last rider, who would then transmit the signal to the others. Stay tuned to see which way McLaren decides to go.
Down with Coffee Lines!
Elsewhere on the floor, Briggo was serving free coffee. Drew Moore, formerly of Dell’s rugged notebook group and now VP of engineering and service at Briggo, showed me the setup. The robot itself is about the size of a small dumpster. Every single move in the coffee-making process has been calibrated and tested a thousand times: how hard to press when snapping on a lid, how long to leave an unclaimed cup on the rack before trashing it, when to refill the coffee beans, milk, sugar, and other ingredients. This barista is fast, efficient, hardly ever makes mistakes, and doesn’t complain or take breaks. Watch out you latte artistes! Soon this thing is going to make pretty foam pictures as well.
There are other nice touches. Rather than storing four kinds of milk, the system stores cream and skim and mixes them according to the order. There’s an app that allows prospective drinkers to specify their tipple and texts them a code when it’s done. No lines. Just elbow to the front, enter your code on the touch panel, and the drink door slides open. Paying customers will get the same great treatment!
These robots can be popped up almost anywhere and generate a lot more revenue per square foot than a Starbucks. The Austin airport (the city is home base) already has one up and going and has another on order. I predict big things for Briggo.