Pit lane walk, or how IaaS technology helps racing teams

    In our blog on Habré, we regularly talk about how IaaS technologies solve the problems of companies from various industries, for example, from the entertainment industry and retail .

    Today we are talking about how cloud technologies help teams win in one of the most spectacular sports - the top car racing series.

    / photo m1try CC

    F1 cars are considered one of the most advanced cars. They are able to accelerate from zero to 300 km / h in 10 seconds and, if necessary, reset the speed from 100 km / h to zero in less than a second. In the turns, the pilots of the cars experience an overload comparable to that of the astronauts upon entering the atmosphere.

    Surprisingly, the teams independently create and design individual components of cars - the engine, chassis, body. The task of engineers and mechanics is to assemble the fastest machine that satisfies the requirements of the annually changing regulations .

    Leading companies from the automotive industry - Mercedes, Ferrari, Honda, Renault - take part in a series of races. Therefore, the degree of competition and the rate of improvement of technology in F1 is very high. To win the title of champion, each of them uses unique solutions and modern technologies. Among these technologies are cloud computing and IaaS.

    How to create a car ...

    As we have said, in the modern "Formula" there are a large number of different requirements of the regulations - hundreds of pages of rules . And here is one of them - teams to test the chassis and hull streamlining are allowed to use the wind tunnel only 25 hours per week. Therefore, F1 engineers paid attention to the hydrodynamic modeling systems (CFD) - their use is not limited in time and tests can be carried out around the clock.

    However, we note that there are some limitations here. To simulate the aerodynamics of a car, teams are allowed to use a maximum of 25 teraflops in calculations in double precision mode (and only on the CPU). Each team provides accurate computing cluster specifications to representatives of the International Automobile Federation (FIA), which is the organizer of the championship.

    For this purpose, the teams “build” their own hardware installations. For example, Renault deployed a computing cluster with 18 thousand cores and Intel Xeon processors. In addition to "physical" computers, teams use cloud computing and IaaS. And for different purposes, for example, for storing large amounts of “aerodynamic data”.

    The Renault team cluster is configured to connect to parallel storage. hownotes the team infrastructure manager Mark Everest (Mark Everest), each compute node has a dedicated connection, so you don't have to spend resources on reading and writing data.

    The teams also use computing power to work with a variety of custom software. These are prototyping tools and production systems, as well as solutions for managing inventory of spare parts and parts for the car, so as not to accidentally arrive at the race with an idle machine.

    Renault Sport uses the Microsoft cloud system to track car assembly processes. It "write" data from the system of planning of corporate resources, information about the changes, as well as the rules of work in the garage. Previously, the team usedfor these purposes, excel-table with a volume of 77 thousand lines. And this despite the fact that F1 racing is considered one of the most technically advanced sports. With the transition to the cloud, each team member was able to quickly track their progress and scope of work.

    ... and win the race

    Team cars are literally crammed with sensors. During the test races, their number is about 400 units, and during the race - about 100 to reduce the weight. These sensors measure almost everything: from basic telemetry (speed, location on a track, etc.) to tire temperature and aerodynamic load.

    “The Formula One car is the true Internet of Things,” says Mercedes Head of IT, Matt Harris. For one racing weekend sensors generate about 300 GB of data.

    / photo m1try CC

    All these data are collected because after the experience of the rider and the technical characteristics of the car, it is the operational telemetry from the track that comes in real time that is responsible for the success of the race. hownoted the former racer António Felix da Costa (António Félix da Costa), without real-time telemetry, “you are passing two circles, and then you stop somewhere on the track, since something is broken”.

    The case of Lewis Hamilton a few years ago illustrates the words of Felix as well as possible. During the race, the Hamilton team engineers radioed him that one of the tires was punctured. He replied that the car is in order, but still drove to the pit stop. Later, they found that the puncture really was, just Hamilton had not yet felt its effect on the behavior of the car.

    To collect and evaluate real-time data, teams resort to using cloud technologies. Once in our blog we toldabout the Sauber team case. At that time, in Sauber F1, using telemetry and track status information, they started simulations in the FlexPod mobile data center (which also sends data from the track to the main office) right during the race. This allowed them to compare the parameters of the virtual model with the current settings of the car and, if possible, make adjustments.

    The FlexPod platform was built on NetApp and Cisco solutions. It combined a two-node cluster NetApp FAS2040 and NetApp SyncMirror. The mobile center (it could easily fit into trucks and airplanes when moving from country to country between stages) included eight Cisco UCS blade servers and a Cisco Nexus switch.

    All Sauber enterprise systems are based on NetApp MetroCluster technology. The infrastructure was virtualized to 90%, and more than fifty VMs were running on the five VMware ESX servers.

    NetApp systems were responsible for storing the results of calculations performed on the HPC cluster. Information was recorded on the clamping force, pressure distribution, air resistance and other processes related to the parameters of the entire car and its components.

    But not only Sauber apply technology to analyze the data of the car right during the race. This system usesMcLaren team. They deployed a cloud infrastructure based on NTT Communications solutions, receiving at their disposal computing resources, data storage and network services. The latter came in handy for the implementation of a driver-in-loop simulator, which with a high accuracy simulates the sensations of driving a real car and helps to create new auto components in the virtual world.

    In the "Formula" value is every thousandth of a second. Even the smallest details are important. Therefore, teams compete not only on the track, but also in the IT field, collecting and analyzing as much data as possible during each race. And cloud technologies help them to achieve optimal "performance" on the highway in the face of fierce competition.

    PS Additional materials from the First Corporate IaaS Blog:

    PPS Several posts from our blog on Habré:

    The main activity of the company IT-GRAD is the provision of cloud services:

    Virtual Infrastructure (IaaS) | PCI DSS Hosting | Cloud FZ-152 | Rent 1C in the cloud

    Also popular now: