#131 2009 · Amazon Web Services · Cloud computing
AWS put its idle servers up for auction rather than let them burn money
the problem
Idle compute perished by the hour
background
EC2 launched in 2006 selling fixed-price, on-demand servers, but a cloud has to be provisioned for its peak load, which means racks sit idle most of the time — capacity that costs the same whether a customer is running on it or not. The standard answers were reserved-capacity discounts or simply eating the waste as the cost of readiness.
On 14 December 2009 AWS opened a second market instead: customers set a maximum bid, ran instances only while that bid cleared a fluctuating Spot Price, and accepted termination on short notice if the price rose past it. Borrowing airline- and commodity-style yield management, AWS let workloads that could tolerate interruption — batch jobs, web crawls, media transcoding — absorb the slack that on-demand pricing left on the table.
the move
Spot Instances: auction idle compute capacity that perishes every hour it sits unused.
the payoff
Became the standard mechanism for monetising slack across cloud providers.
what came after
Google later offered its own reduced-price preemptible tier on the same interruptible-discount structure, and spot markets are now a standard way large-scale batch and ML workloads buy cheap cycles. The "auction" itself did not survive: in November 2017 Amazon replaced the sealed-bid market entirely with a retail price it sets and does not disclose the algorithm for, and researchers who compared prices before and after found the change was accompanied by a broad price increase — averaging roughly 37% to 61% across AWS's North American regions — not the drop a pure reliability cut should have produced.
filed under
references
- [1]Amazon Web Services Announces Spot Instances for Amazon EC2Amazon (press release), 2009press.aboutamazon.com
- [2]Analyzing AWS Spot Instance PricingIEEE IC2E / UC Santa Barbara, 2019sites.cs.ucsb.edu