1. A Case Study on How Economic Frameworks Can Bail Out Systems Research Emin Gün Sirer Ryan Peterson, Bernard Wong Department of Computer Science, Cornell University United Networks, LLC September 3, 2009

2. Problem Domain What is the most efficient way to disseminate a large set of files to a large set of clients?

3. Client-Server server clientsInefficient High cost of ownership

4. Peer-to-Peer peer Limited information No control or performance guarantees

5. Traditional Systems Research Complex problems lead to complex protocols with many parameters Parameters picked based on a combination of gut feeling and experimentation Systems tailored to specific workloads Only intellectual honesty keeps researchers from perfect performance on benchmarks! Novelty by aggregation Pick k out of n: CDNs, network coordinates, geolocation, network tomography, …

6. BitTorrent Robust, resilient protocol A combination of randomization and heuristics Two years until characterized as auctions based on received bandwidth Lots of incremental research! Every 10-line tweak => 1 paper Hacks: Share ratios, manual pruning, … We lack the tools to tackle difficult problems

7. Antfarm Goals High performance Low cost of deployment Performance guarantees Administrator control over swarm performance Accounting Enables different resource contribution policies

8. Antfarm Approach Key insight: view content distribution as an optimization problem Hybrid architecture P2P swarming with a logically centralized coordinator Clean slate protocol

9. Antfarm System Model coordinator seeder altruist

10. Strawman Coordinator One could schedule every data transfer in the system All packets for all time Unscalable, impractical! Antfarm coordinator makes critical decisions based on observed dynamics

11. Antfarm Coordinator Models swarm dynamics Measures and extracts key parameters Formulates optimization problem Calculates optimal bandwidth allocation Enacts allocation decisions Maximizes aggregate bandwidth Minimizes average download time

12. Antfarm Formalization Maximize system-wide aggregate bandwidth subject to a bandwidth constraint

13. Response Curves slope = 1 slope = 0

14. Response Curves Swarm aggregate bandwidth (KB/s) 1500 1000 500 0 Seeder bandwidth (KB/s) 0 257510050

15. Swarm Dynamics Swarms exhibit different dynamics based on size, peer resources, network conditions...

16. Antfarm Optimization σAσA σCσC σBσB σ A + σ B + σ C = B A B C

17. Performance Control Can provide swarm performance guarantees Guarantee minimum level of service Prioritize swarms

18. Antfarm Allocation σA′σA′σC′σC′σB′σB′ σ A ′+ σ B ′+ σ C ′= B A B C

19. Problems The optimization problem has limitations Requires knowing total bandwidth constraint Applies only to full caches under the control of the coordinator At the wire protocol level, there are many opportunities for malfeasance

20. Antfarm++ Use a currency scheme, called karma, to keep track of peers’ resource consumption and contribution Every exchange costs some karma to the peer receiving a block, brings karma to the peer providing the block and bandwidth

21. peer A purseledger Wire Protocol Coordinator mints small, unforgeable tokens Peers trade each other tokens for blocks Peers return spent tokens to the coordinator as proof of contribution coordinator peer B purseledger

22. Tokens & Exchange Rates Every swarm uses a per-swarm currency, every token has source-destinations marked Token flows reveal the internal network dynamics to the coordinator The coordinator can provide incentives in the form of exchange rates Exchange rates encode the differential in the desirability of participating in different swarms => no need to estimate total BW

23. Benefits of a Virtual Currency The cost and benefit of a transaction need not be symmetrical A block always costs 1 token to the buyer i A block can bring γ i karma to the seller Provide an incentive for peers to join and seed swarms in need Provide an incentive for peers to join and seed swarms in need

24. Antfarm Performance

25. Swarm Starvation BitTorrent starves the singleton swarm

26. BitTorrent: Starves New Swarm Swarms, ordered largest to smallest newself-sufficientsingleton

27. Antfarm: Seeds New Swarm Swarms, ordered largest to smallest newself-sufficientsingleton

28. Conclusions Hard problems in systems research can benefit immensely from the application of economics http://flixq.com showcases our current work http://flixq.com

29. Related Work Content Distribution Networks Akamai, CoBlitz, CoDeeN, ECHOS, Coral, Slurpie, YouTube, Hulu, GridCast, Tribler, Joost, Huang et al. 2008,... P2P Swarming BitTorrent, BitTyrant, PropShare, BitTornado, BASS, Annapureddy et al. 2007, Guo et al. 2005,... Incentives and microcurrencies Dandelion, BAR Gossip, Samsara, Karma, SHARP, PPay, Kash et al. 2007,...

30. Questions?

31. PlanetLab Deployment

33. Results High performance out-of-the-box Discovers which swarms make the most efficient use of seeder bandwidth Automatically prioritizes new swarms over self-sufficient swarms

34. Antfarm System Model coordinator cache altruist multi-swarm peer datacenter