260 likes | 344 Views
Employing Agent-based Models to study Interdomain Network F ormation, Dynamics & Economics . Aemen Lodhi (Georgia Tech). Workshop on Internet Topology & Economics (WITE’12). Outline. Agent-based modeling for AS-level Internet Our model: GENESIS Application of GENESIS
E N D
Employing Agent-based Models to study Interdomain Network Formation, Dynamics & Economics AemenLodhi(Georgia Tech) Workshop on Internet Topology & Economics (WITE’12)
Outline • Agent-based modeling for AS-level Internet • Our model: GENESIS • Application of GENESIS • Large-scale adoption of Open peering strategy • Conclusion
What is the environment that we are we trying to model? • Autonomous System level Internet • Economic network Internet Transit Provider Transit Provider Enterprise customer Content Provider Content Provider Enterprise customer
What is the environment that we are we trying to model? • Complex, dynamic environment • Mergers, acquisitions, new entrants, bankruptcies • Changing prices, traffic matrix, geographic expansion • Co-evolutionary network • Self-organization • Information “fuzziness” • Social aspects: 99% of all peering relationships are “handshake” agreements* *”Survey of Characteristics of Internet Carrier Interconnection Agreements 2011” – Packet Clearing House
What are we asking? • Aggregate behavior • Is the network stable? • Is their gravitation towards a particular behavior e.g., Open peering • Is their competition in the market? • Not so academic questions • Is this the right peering strategy for me? • What if I depeerAS X? • Should I establish presence at IXP Y? • CDN: Where should I place my caches?
Different approaches • Analytical / Game-theoretic approach • Empirical studies • Generative models e.g., Preferential attachment • Distributed optimization • Agent-based modeling
Why to use agent-based modeling? • Incorporation of real-world constraints • Non-uniform traffic matrix • Complex geographic co-location patterns • Multiple dynamic prices per AS • Different peering strategies at different locations • Scale – hundreds of agents • What-if scenarios • Understanding the “process” and not just the “end-state”
Why not to use agent-based modeling? • Large parameterization space • Systematic investigation of full parameter space is difficult • Validation • Computational cost • Under some circumstances reasoning may be difficulte.g. instability in a model with hundreds of agents
The model: GENESIS* • Agent based interdomain network formation model • Fundamental unit: An agent (AS) with economic interests • Incorporates • Co-location constraints in provider/peer selection • Traffic matrix • Public & Private peering • Set of peering strategies • Peering costs, Transit costs, Transit revenue *AemenLodhi, AmoghDhamdhere, Constantine Dovrolis, “GENESIS: An agent-based model of interdomain network formation, traffic flow and economics,” InfoCom 2012
Peering link at top tier possible across regions Geographic presence & constraints Geographic overlap Link formation across geography not possible Regions corresponding to unique IXPs
The model: GENESIS* Fitness = Transit Revenue – Transit Cost – Peering cost • Objective: Maximize economic fitness • Optimize connectivity through peer and transit provider selection • Choose the peering strategy that maximizes fitness
Peering strategies • Restrictive: Peer only to avoid network partitioning • Selective: Peer with ASes of similar size • Open: Every co-located AS except customers • Choose peering strategy that is predicted to give maximum fitness
Peering strategy adoption Open Selective Open • No coordination, limited foresight • Eventual fitness can be different • Stubs always use Open peering strategy 1 2 3 Time Transit Provider selection Depeering Peering
Application of GENESIS:Analysis of peering strategy adoption by transit providers in the Internet* *AemenLodhi, AmoghDhamdhere, Constantine Dovrolis, “Analysis of peering strategy adoption by transit providers in the Internet,” NetEcon 2012
Motivation: Existing peering environment • Increasing fraction of interdomain traffic flows over peering links* • How are transit providers responding? Transit Provider Access ISP/Eyeballs Content Provider/CDN *C. Labovitz, S. Iekel Johnson, D. McPherson, J. Oberheide and F. Jahanian, “Internet Interdomain Traffic,” in ACM SIGCOMM, 2010
Motivation: Existing peering environment • Peering strategies of ASes in the Internet (source: PeeringDBwww.peeringdb.com) • Transit Providers peering openly ?
Approach • Agent based computational modeling • Corroboration by PeeringDB data • Scenarios *Stubs always use Open Without-open • Selective • Restrictive With-open • Selective • Restrictive • Open vs.
Collective impact of Open peering on fitness of transit providers • Cumulative fitness reduced in all simulations
Impact on fitness of individual transit providers switching from Selective to Open • 70% providers have their fitness reduced
Why do transit providers adopt Open peering? • Affects: • Transit Cost • Transit Revenue • Peering Cost v Save transit costs x y But your customers are doing the same! z w
Why gravitate towards Open peering? x adopts Open peering x regains lost transit revenue partially x lost transit revenue Options for x? x y Not isolated decisions Network effects !! Y peering openly z w, traffic passes through x again! z w z w, z y, traffic bypasses x
Conclusion • Employ agent-based models for large-scale study of interdomain network formation • Parameterization and validation are difficult • Agent-based models can reveal surprising behavior
Conclusion • Gravitation towards Open peering is a network effect for transit providers (70% adopt Open peering) • Economically motivated strategy selection • Myopic decisions • Lack of coordination • Extensive Open peering by transit providers in the network results in collective loss