Keynote Playing games in the clouds

on YouTube ↗29m

" summary

Uses Nash bargaining solution to compute exact allocations (example: two machines handling four jobs -> 3.5 jobs to A, 0.5 to B, measured response time 1.33s vs 1.39/1.36 for nearby reallocations), and shows second-price auction makes truthful reporting a dominant strategy for machine capabilities. Implements an auctioneer in a cloud foundry–style scheduler where VMs bid with memory/disk, experiments show the bargaining-based allocator equalizes per-job response time and drops too-slow servers rather than overloading them.

AI-written summary · about

" chapters (6)

1. 0:00Start
2. 3:49WHAT IS A GAME?
3. 15:42HOW CAN WE MODEL THIS?
4. 18:13THE SOLUTION
5. 21:34ALLOCATION ALGORITHM
6. 23:15APP PLACEMENT IN CLOUD FOUNDRY

" related talks

• 52mDashboard Mechanisms for Online MarketplacesJason Hartline · Information Design and Data Science 2019
• 44mOnline Advertisements and Online AlgorithmsNikhil Devanur · Algorithmic Game Theory and Practice 2015
• 23mKeynote: gRPC, A Foundation for AIMuninder Sambi, Arunkumar Jayaraman · gRPConf 2025
• 08mMaximizing ROI Through Distributed Computing and Multi-CloudRavi Kumar Amaresam · Conf42 Machine Learning 2025
• 11mKeynote: Multicluster Strategies to “Kubernetify” Legacy AppsSai Vennam · Kubernetes Forum Seoul 2019
• 23mKeynote: Cloud Native Java, ReduxJosh Long · J-Fall Virtual 2020
• 39mBuild blazingly fast, ultra lightweight cloud native Java enterprise applicationsRăzvan Prichici · Voxxed Days Cluj-Napoca 2019
• 42mPredictive Load-Balancing: Unfair but Faster & more RobustSteve Gury · Strange Loop 2017