Person:

Ng, Chaki

The Egg project provides a vision and implementation of how heterogeneous computational requirements will be supported within a single grid and a compelling reason to explain why computational grids will thrive. Environment computing, which allows a user to specify properties that a compute environment must satisfy in order to support the user’s computation, provides a how. Economic principles, allowing resource owners, users, and other stakeholders to make value and policy statements, provides a why. The Egg project introduces a language for defining software environments (egg shell), a general type for grid objects (the cache), and a currency (the egg). The Egg platform resembles an economically driven Internetwide Unix system with egg shell playing the role of a scripting language and caches playing the role of a global file system, including an initial collection of devices.

The widespread deployment of high-speed internet access is ushering in a new era of distributed computing, in which parties both contribute to a global pool of shared resources and access the pooled resources to support their own computing needs. We argue that system designers must explicitly address the self-interest of individual parties if these next-generation computing systems are to flourish. We propose strategyproof computing, a vision for an open computing infrastructure in which resource allocation and negotiation schemes are incentive-compatible, and individual parties can treat other resources as their own. In this paper we outline key guiding principles for the vision of strategyproof computing, define the strategyproof computing paradigm, and lay out a systems-related research agenda.

We consider the problem of designing fast and strategyproof exchanges for dynamic resource allocation problems in distributed systems. The exchange is implemented as a sequence of auctions, with dynamically arriving requests from agents matched with each auction. Each auction is associated with some consignment of the resources from a single seller. We provide a simple Virtual Worlds (VW) construction, that extends a fast and strategyproof mechanism for a single auction to apply to this sequence-of-auctions setting. Rather than match each buyer with a single auction, the VW mechanism allows buyers to be considered for multiple auctions while retaining strategyproofness.