0101011010100111111111010001010011011010 0010101001110101001000100100010101110111 0101000101111110010111000001101010000111 0111111001000010101000111000101010100100 1110110000100000110010000010111101000100 0011100000000110000100001011010111100111 1111010010010110111110010001000010001101 0110001011111011110010011101010100010101 0100010011010001101101110001000000000110 0110001111001001001011000101011001110000 0111011000011001000110011001111001010100 1010100000110100101000110010110010110111 0100111100011111001001010111001000111101 0000011011010010011100001111110010110110 0110011001100010111111101111100011000010 1001010111111101011001001010001011110110 0111100010110011110000011110011010111101 1101101100110101100110111110010000000111 0111001000001100110111111001101111100011

Jeffrey Shneidman Maxwell Dworkin 121 33 Oxford Street Harvard University, 02138 Best Contact Phone (Mobile): Graduate student in Systems Research at Harvard and the Economics and Computer Science research groups.

Ordered by subject, and then reverse chronological.

• Why Markets Could (But Don't Currently) Solve Resource Allocation Problems in Systems
  Jeffrey Shneidman, Chaki Ng, David C. Parkes, Alvin AuYoung, Alex C. Snoeren, Amin Vahdat, and Brent Chun.
  (First published in: Proceedings of Tenth Workshop on Hot Topics in Operating Systems) [HotOS-X 2005]
  Position Paper: This paper has three goals. The first goal is to explore why markets can be appropriate to use for allocation, when simpler allocation mechanisms exist. The second goal is to demonstrate why a new look at markets for allocation could be timely, and not a re-hash of previous research. The third goal is to point out some of the thorny problems inherent in market deployment and to suggest action items both for market designers and for the greater research community.
  [PDF] [PDF of HotOS talk]

• ICE: An Iterative Combinatorial Exchange
  David C. Parkes, Ruggiero Cavallo, Nick Elprin, Adam Juda, Sebastien Lahaie, Benjamin Lubin, Loizos Michael, Jeffrey Shneidman, Hassan Sultan.
  (ACM Conference on Electronic Commerce) [ACM EC 2005]
  We present the first design for a fully expressive iterative combinatorial exchange. ICE can allocate resources among mixed buyers and sellers (e.g.: users that simultaneously buy or sell bundles of items). The iterative nature allows a user to be inexact in their initial valuation (often desirable in the real world) and reduces the communication and computation complexity of the mechanism. All features in this paper have been fully implemented in software.
  [PDF]

• TBBL: A Tree-Based Bidding Language for Iterative Combinatorial Exchanges
  Ruggiero Cavallo, David C. Parkes, Adam Juda, Adam Kirsch, Alex Kulesza, Sebastien Lahaie, Benjamin Lubin, Loizos Michael, and Jeffrey Shneidman.
  (International Joint Conferences on A.I. 2005: Workshop on Advances in Preference Handling) [IJCAI 2005 Preference Workshop]
  We present a novel tree-based logical bidding language, TBBL, for preference elicitation in combinatorial exchanges (CEs). TBBL provides new expressiveness for two-sided markets with agents that are both buying and selling goods. Moreover, the rich semantics of TBBL allow the language to capture new structure, making it exponentially more concise than OR* and L_GB for preferences that are realistic in important domains for CEs. With simple extensions TBBL can subsume these earlier languages. TBBL can also explicitly represent partial information about valuations. The language is designed such that the structure in TBBL bids can be concisely captured directly in mixed-integer programs for the allocation problem. We illustrate TBBL through examples drawn from domains to which it can be (and is being) applied, and motivate further extensions we are currently pursuing.
  [PDF]
  

• Mirage: A Microeconomic Resource Allocation System for SensorNet Testbeds.
  Brent N. Chun, Philip Buonadonna, Alvin AuYoung, Chaki Ng, David C. Parkes, Jeffrey Shneidman, Alex C. Snoeren, and Amin Vahdat.
  (Proceedings of the 2nd IEEE Workshop on Embedded Networked Sensors, May 2005.) [EMNETS 2005]
  In this paper, we argue that a microeconomic resource allocation scheme, specifically the combinatorial auction, is well suited to testbed resource management.
  [PDF]

• Also available on request is the not-yet-published conference paper, Handling Rational Failure in Distributed Consensus, 2007..

• Faithfulness in Internet Algorithms, Jeffrey Shneidman and David C. Parkes and Laurent Massoulié
  (ACM SIGCOMM 2004 Workshop on Practice and Theory of Incentives and Game Theory in Networked Systems) [PINS 2004]
  A more applied, less formal consideration of faithfulness as defined in the PODC04 paper. Whereas PODC04 demonstrates a way to prove faithfuless, this work seeks to disprove faithfulness in existing systems by finding a contradiction using a backtracing technique. In this paper, we analyze the BitTorrent protocol/system. Presents open questions about dealing with faulty nodes (in the classic distributed systems sense: Byzantine, failstop, etc.) in rational settings.
  [PDF] [PS] [Talk Slides (PDF)] [BibTex Entry]

• Specification Faithfulness in Networks with Rational Nodes, Jeffrey Shneidman and David C. Parkes
  (Twenty-Third Annual ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing) [PODC 2004]
  In this paper, we present a framework for proving specification faithfulness given certain knowledge assumptions. In other words, we start out by saying, "It is useful to prove that an implementation correctly implements a specification. But even if a designer creates and publishes a proveably correct implementation, how can one prove that this implementation will be followed by nodes in a network (of rational nodes)?" This problem boils down to proving the incentive-, algorithm-, and communication-compatibility of a specification. This in turn can be used to prove the specification robust to rational mantipulation or "rational failure" -- in other words, all rational nodes will be faithful (for some definition of rational, which we formalize in the paper.)
  [PDF] [BibTex Entry]

• Distributed Implementations of Vickrey-Clarke-Groves Mechanisms, David C. Parkes and Jeffrey Shneidman
  (3rd Int. Joint Conf. on Autonomous Agents and Multi Agent Systems) [AAMAS 2004]
  One drawback of mechanism design (MD) is that computation is implicitly centralized, with a central planner taking all decisions. This paper considers distributed implementations, in which the outcome is determined by the self-interested agents themselves. Clearly this introduces new opportunities for manipulation. We propose a number of principles to guide the distribution of computation, focusing in particular on Vickrey-Clarke-Groves mechanisms for implementing outcomes that maximize total value across agents. Our solutions bring the complete implementation into an ex post Nash equilibrium.
  [PDF] [BibTex Entry]

• Using Redundancy to Improve Robustness of Distributed Mechanism Implementations, Jeffrey Shneidman and David C. Parkes
  (Poster paper in 2003 ACM Conference on E-commerce [ACM EC03]
  An early version of the PODC04 work.
  [PDF (2-page Poster version)] [PS (2-page Poster version)] [BibTex Entry]

• Rationality and Self-Interest in Peer to Peer Networks, Jeffrey Shneidman and David C. Parkes
  (2nd International Workshop on Peer-to-Peer Systems [IPTPS 2003]
  Introduces mechanism design (MD) as a tool to the IPTPS crowd; presents open questions that arise when using MD in peer to peer settings.
  [PDF] [PS] [Talk Slides (HTML)] [Practice Talk (MP3)] [BibTex Entry]

• Network-Aware Operator Placement for Stream-Processing Systems
  Peter Pietzuch, Jonathan Ledlie, Jeffrey Shneidman, Mema Roussopoulos, Matt Welsh, and Margo Seltzer.
  (Proceedings of the 22nd International Conference on Data Engineering) [ICDE 2006]
  This paper describes a stream-based overlay network (SBON), a layer between a stream-processing system and the physical network that manages operator placement for stream-processing systems. Our design is based on a cost space, an abstract representation of the network and on-going streams, which permits decentralized, large-scale multi-query optimization decisions. We present an evaluation of the SBON approach through simulation, experiments on PlanetLab, and an integration with Borealis, an existing stream-processing engine. Our results show that an SBON consistently improves network utilization, provides low stream latency, and enables dynamic optimization at low engineering cost.
  [PDF]

• Optimizing Streaming Applications with Self-Interested Users using MDPOP
  Boi Faltings, David Parkes, Adrian Petcu, and Jeffrey Shneidman
  (International Workshop on Computational Social Choice) [COMSOC 2006]
  In this paper we deal with the problem of optimally placing a set of query operators in an overlay network. However, like my work in faithfulness, these users and nodes are self-interested. Each user has private value information about the operators associated with a query and about the utility from different combinations of operator placements. Each server in the overlay network is able to perform some set of operators, and servers differ in their network and computational characteristics.
  [PDF]

• A Cost-Space Approach to Distributed Query Optimization in Stream Based Overlays,
  Jeffrey Shneidman, Peter Pietzuch, Matt Welsh, Margo Seltzer, and Mema Roussopoulos
  (Proceedings of the 1st IEEE International Workshop on Networking Meets Databases) [NetDB 2005]
  Distributed stream-based applications, such as continuous query systems, have network scale and time characteristics that challenge traditional distributed query optimization. The optimization sub-problems of plan generation and service placement should be integrated to meet these challenges. These tasks have typically been treated as independent sub-problems because of the complexity of their integration. We suggest cost spaces as one way to mitigate this complexity. We further consider how cost spaces can be used to allow tractable multi-query optimization.
  [PDF]

• Evaluating DHT-Based Service Placement for Stream-Based Overlays,
  Peter Pietzuch, Jeffrey Shneidman and Matt Welsh, Margo Seltzer, and Mema Roussopoulos
  (Proceedings of the 4th International Workshop on Peer-to-Peer Systems) [IPTPS 2005]
  Recent work on operator placement proposes to leverage routing paths in a distributed hash table (DHT) to obtain a set of candidate nodes for service placement. We evaluate the appropriateness of using DHT routing paths for service placement in an stream-based overlay network, when aiming to minimize network usage.
  [PDF] [PPT]

• Open Problems in Data Collection Networks,
  Jonathan Ledlie, Jeffrey Shneidman, Matt Welsh, Mema Roussopoulos, and Margo Seltzer
  (11th ACM SIGOPS European Workshop) [SIGOPS EW 2004]
  Proposes data collection network as answer to question of how to create "vision" applications. Identifies topics not handled well by the union of research areas of sensor networks, continuous query, peer to peer. Presents overview of current Hourglass project.
  [PDF]

• Scooped, Again, Jonathan Ledlie and Jeffrey Shneidman and Margo Seltzer and John Huth
  (2nd International Workshop on Peer-to-Peer Systems [IPTPS 2003]
  Position paper: Peer to peer research and scientific computing research are solving the same problems and should work together.
  [PDF] [PS]

• Path Optimization in Stream-Based Overlay Networks,
  Peter Pietzuch, Jeffrey Shneidman, Mema Roussopoulos, Matt Welsh, and Margo Seltzer
  (Harvard Technical Report TR-26-04) Fall, 2004
  In this paper, we describe a novel, network-aware path optimization algorithm for stream-based overlay networks. Our approach is based on a spring relaxation model that operates in a metric space defined by the pairwise node latency within the SBON. This relaxation placement algorithm utilizes the underlying network resources efficiently, is capable of performing inter-stream optimization, and can be implemented in a scalable and decentralized way.
  [PDF]

• Collecting Data for One Hundred Years, Jeffrey Shneidman and Bryan Choi and Margo Seltzer (Harvard Technical Report TR-06-03)
  What would change about network design if they were meant to run for 100 years without aid from a system administrator? An precursor to the Hourglass project.
  [PDF]

• Chip: A Temporary, Memory Based Network Filesystem (with Chambers, Steinbach) (6.824 at MIT), Spring 2001.

• Six ways from Sunday: A Security Audit of Harvard's Crimson Cash Debit System (with Wilken, Su) (CS222), Spring 2000.

• Depth-Scaled Biased Stochastic Sampling (with Enos, Ruml) (CS182), Fall 1999.