In motorsport racing, Formula One (F1) and IndyCar provide a great experience for spectators, an impressive safety record for the drivers considering the nature of the sport, and great returns for their sponsors. They also have evolved different approaches to solving their technical challenges, in particular how they spread their costs across the teams participating in each sport. There’s a close correlation between the divergent approaches taken by F1 and IndyCar, and the strategies we use in selecting infrastructure for enterprise datacenters. I’ll be diving deeper into this topic at Cisco Live! 2017 in Berlin later this month (details below), but read along and give me feedback if you’d like a preview.
Formula One cars are individually designed, hand assembled by each manufacturer and precisely tailored to the driver, the engine and the powertrain. They are the pinnacle of optimization. They also carry a price tag appropriate for that optimization, with many teams spending upwards of $100M each year just on the R&D and production of the cars.
All teams in IndyCar racing use the same chassis and running gear, one of two engine options, and for most teams, aero packs supplied by the engine manufacturer. They are as fast on the straight and banked curves as F1 cars, but unlike F1 cars, they cannot maintain that speed around the flat corners and chicanes common across all F1 tracks. The selection of a single chassis and two engines allow costs of R&D and manufacturing to be shared across all teams, resulting in a typical IndyCar team spending a mere $3M each year on their car, about 3% of what an equivalent F1 team will spend.
What works in motorsports also works in the data center. Some applications benefit from tight optimization between the application and infrastructure, and it is both practical and a good investment to spend a little more for that optimized approach. Other application environments, particularly those providing more of a cloud-centric capability, don’t realize any meaningful benefit from optimizing for a particular goal. They are better deployed using Engineered Systems, which includes the Dell EMC Vxblock System, Dell EMC VxRack and VxRail, where the cost of R&D and manufacture is amortized across all users.
Imagine a system that arrives in your data center, ready to go, with ACI provisioned on the switches, the compute blades and the hypervisor; a system where the network traffic is appropriately and correctly distributed across all switch port groups; a system where the storage traffic and the production network traffic are balanced, and where licenses for network and SAN ports are not a sudden surprise.
There are other clear benefits from integrating all components into a system. The sunroof and the radio in many modern cars are fully integrated into the operational aspect of the car. Answer a call on your Bluetooth equipped cellphone, and the radio mutes, while the speakers become a part of the phone audio system; press the lock button on the remote, and the sunroof automatically closes.
The network is a critical component for all services in the data center, and yet too frequently, it is treated as an after-thought. Engineered Systems design and build-in the network as part of the overall system in the same way that the radio and sunroof are part of the system in modern cars. The network is integrated as part of the design process, not as an optional component, and the rest of the system benefits from that integration. That means you get to benefit as well.
These benefits extend into the day to day operational aspects. We’ve all muttered “there but for the grace of god” as we read horror stories about supposedly redundant networks that turned out to be anything but redundant. As each engineered system is, like the IndyCar chassis, a manufactured object rather than an ever-evolving prototype like the F1 chassis, the redundancy designed-in to an Engineered System will always exist when those systems are operating in the data center.
There is a place for highly optimized environments in the data center, where the performance of the application demands that optimization, and where the benefit justifies the incremental cost. Many applications however, don’t need, and provide no incremental benefits from over optimizing. For those environments, you can benefit from the reduced costs and integrated network that Engineered Systems provide. Dell EMC is unique in offering a selection of products from servers, network switches, storage arrays and software through to fully Engineered Systems. A single supplier can now meet your needs for optimized environments and for general purpose, re-usable infrastructure.
If you are planning on visiting Cisco Live! in Berlin later this month, I’d like to invite you to a session I’ll be co-hosting with Stefan Daiber from ScanPlus on “Letting software drive your data center foundation services”, where we will be discussing this topic in greater depth (Wednesday, February 22 at 12:30 CET in the Innovation Theater, Hall 3.2).
Sources for comparisons: