Race weekscene 9 / 26~2 min
Other people's computers
Friday. The building, located at last.
Since Act 0, this story has said the same sentence four ways: a building Case will never visit, a machine Case will never see, a place nobody at Traversal owns, somewhere far away holding the entry. Friday of race week, the building finally gets its chapter. It has racks.
It is a data centerConcept · lights on your mapdata centerA facility built to house and run computing equipment: racks of servers, storage and networking, engineered around electricity, cooling, connectivity and physical security. The building around the computer is part of the system.: a warehouse of racks engineered around four unglamorous things, power, cooling, connectivity, and locked doors. The machines Traversal rents live in several of them. Nobody at Traversal has keys. Nobody at Traversal wants keys, and the not-wanting has a name.
The ownership model, not the absence of computers
cloud computingConcept · lights on your mapcloud computingComputing resources and managed services delivered over a network from provider-run data centres, created on demand and billed by usage. Not the absence of computers: a different ownership model for them. The major providers are AWS, Microsoft Azure, and Google Cloud. is the arrangement this story has been living inside without naming: the computers exist, intensely, in those buildings, and Traversal creates them with an API call and pays by the hour. The cloud is somebody else’s computers plus, and the plus is the point: on-demand creation, metered billing, and the whole catalogue of managed machinery the next scenes tour. When Sunday needs thirty machines for four hours, Traversal will summon them, use them, and dismiss them, and the meter will read accordingly.
The opposite arrangement is standing in a tent at the finish line. on-premisesConcept · lights on your mapon-premisesHardware an organisation runs primarily for its own use, in its building or its own racks. The alternative to renting on demand, and the practical question is which fits each workload. means owning and running the hardware yourself, and the timing crew’s laptops are the genuine article: bought, carried, babied, backed up by the crew chief personally. Not a museum piece, a correct choice. A finish line needs machines that work with no network at all, and you cannot rent what has to sit on a folding table in the rain. Cloud where elasticity matters, on-prem where the folding table does. The fluency is knowing it was a choice both times.
So what, exactly, does Traversal summon with an API call? A machine that is not quite a machine. Next scene.
End of scene
This scene covers: data center, cloud computing, on-premises