Peer-to-peer sharing systems are becoming increasingly popular and an exciting new class of innovative, internet-based data management systems. In these systems, users contribute their own resources (processing units and storage devices) and content (i.e., documents) to the P2P community. We focus on the management of content and resources in such systems. Our goal is to harness all available resources in the P2P network so that the users can access all available content efficiently. Efficiency is taken both from (i) the point of view of the system, in that we strive to ensure fair load distribution among all peer nodes, and (ii) from the point of view of the users, in that we strive to ensure low user-request response times.

Desktop grids have recently been used to perform some of the largest computations in the world and have the potential to grow by several more orders of magnitude. However, current approaches to utilizing desktop resources require either centralized servers or extensive knowledge of the underlying system, limiting their scalability.

Abstract — We propose a biologically inspired and fullydecentralized approach to the organization of computation that is based on the autonomous scheduling of strongly mobile agents on a peer-to-peer network. Our approach achieves the following design objectives: near-zero knowledge of network topology, zero knowledge of system status, autonomous scheduling, distributed computation, lack of specialized nodes. Every node is equally responsible for scheduling and computation, both of which are performed with practically no information about the system. We believe that this model is ideally suited for large-scale unstructured grids such as desktop grids. This model avoids the extensive system knowledge requirements of traditional Grid scheduling approaches. Contrary to the popular master/worker organization of current desktop grids, our approach does not rely on specialized super-servers or on application-specific clients. By encapsulating computation and scheduling behavior into mobile agents, we decouple both application code and scheduling functionality from the underlying infrastructure. The resulting system is one where every node can start a large grid job, and where the computation naturally organizes itself around available resources. Through the careful design of agent behavior, the resulting global organization of the computation can be customized for different classes of applications. In a previous paper, we described a proof-of-concept prototype for an independent task application. In this paper, we generalize the scheduling framework and demonstrate that our approach is applicable to a computation with a highly synchronous communication pattern, namely Cannon’s matrix multiplication. I.

Abstract—Designers face many system optimization problems when building distributed systems. Traditionally, designers have relied on optimization techniques that require either prior knowledge or centrally managed runtime knowledge of the system’s environment, but such techniques are not viable in dynamic networks where topology, resource, and node availability are subject to frequent and unpredictable change. To address this problem, we propose collaborative reinforcement learning (CRL) as a technique that enables groups of reinforcement learning agents to solve system optimization problems online in dynamic, decentralized networks. We evaluate an implementation of CRL in a routing protocol for mobile ad hoc networks, called SAMPLE. Simulation results show how feedback in the selection of links by routing agents enables SAMPLE to adapt and optimize its routing behavior to varying network conditions and properties, resulting in optimization of network throughput. In the experiments, SAMPLE displays emergent properties such as traffic flows that exploit stable routes and reroute around areas of wireless interference or congestion. SAMPLE is an example of a complex adaptive distributed system. Index Terms—Feedback, learning systems, mobile ad hoc network routing. I.

The complexity and dynamism of modern network raise several challenges in the design and development of communication services. The unbearable costs in configuration and management call for autonomic approaches, in which services are able to selfconfigure and self-adapt their activities without human intervention. The need for ubiquity of service provisioning calls for the capability of services of adapting their behavior depending on the current situation (social and spatial) in which they are used. In this paper, after having discussed the need for innovative approaches facilitating the design, development, and execution of autonomic and situation-aware services, we try to analyze the key features that should underlies such a general approach, proposes a general-purpose architecture centered around the abstraction of “agent communication elements”, and sketch the main research thrusts that should be pursued for the realization of the vision. 1. The Vision The Internet as we know it today will have to become like an immense ecology of composite, highly distributed, pervasive, communication-intensive services [KepC03, Zam05]. Such services should be able to: (i) autonomously detect and organize the knowledge necessary to understand the general context – physical, technological, social, user-specific and request-specific – in which they operate; (ii) self-adapt and self-configure their functioning to get the best from any situation, so as to meet the needs of diverse users in diverse situation without explicit human intervention. These features will enable a wide range of new activities that are simply not possible or

Counting in general, and estimating the cardinality of (multi-) sets in particular, is highly desirable for a large variety of applications, representing a foundational block for the efficient deployment and access of emerging internetscale information systems. Examples of such applications range from optimizing query access plans in internet-scale databases, to evaluating the significance (rank/score) of various data items in information retrieval applications. The key constraints that any acceptable solution must satisfy are: (i) efficiency: the number of nodes that need be contacted for counting purposes must be small in order to enjoy small latency and bandwidth requirements; (ii) scalability, seemingly contradicting the efficiency goal: arbitrarily large numbers of nodes nay need to add elements to a (multi-) set, which dictates the need for a highly distributed solution, avoiding server-based scalability, bottleneck, and availability problems; (iii) access and storage load balancing: counting and related overhead chores should be distributed fairly to the nodes of the network; (iv) accuracy: tunable, robust (in the presence of dynamics and failures) and highly accurate cardinality estimation; (v) simplicity and ease of integration: special, solution-specific indexing structures should be avoided. In this paper, first we contribute a highly-distributed, scalable, efficient, and accurate (multi-) set cardinality estimator. Subsequently, we show how to use our solution to build and maintain histograms, which have been a basic building block for query optimization for centralized databases, facilitating their porting into the realm of internet-scale data networks. 1

Abstract—In this paper, we present a macroscopic characterization of agent-based load balancing in homogeneous minigrid environments. The agent-based load balancing is regarded as agent distribution from a macroscopic point of view. We study two quantities on minigrids: the number and size of teams where agents (tasks) queue. In macroscopic modeling, the load balancing mechanism is characterized using differential equations. We show that the load balancing we concern always converges to a steady state. Furthermore, we show that load balancing with different initial distributions converges to the same steady state gradually. Also, we prove that the steady state becomes an even distribution if and only if agents have complete knowledge about agent teams on minigrids. Utility gains and efficiency are introduced to measure the quality of load balancing. Through numerical simulations, we discuss the utility gains and efficiency of load balancing in different cases and give a series of analysis. In order to maximize the utility gain and the efficiency, we theoretically discuss the optimization of agents ’ strategies. Finally, in order to validate our proposed agentbased load balancing mechanism, we develop a computing platform, called Simulation System for Grid Task Distribution (SSGTD). Through experimentation, we note that our experimental results in general confirm our theoretical proofs and numerical simulation results on the proposed equation system. In addition, we find a very interesting phenomenon, that is, our agent-based load balancing mechanism is topology-independent.

We consider the problem of deploying and managing federated services that run on federated systems spanning multiple collaborative organizations. In particular, we present a peer-to-peer framework targeted to the construction of self-managing services that automatically adjust the number of service components and their placements in response to changes in the system or client loads. Our framework is completely decentralized, depending only on a modest amount of loosely synchronized global state. More specifically, our framework is comprised of a set of per-node monitoring agents and per-service-component management agents that periodically exchange information about the state of the system and of the service with each other using a gossiping protocol. Each management agent then periodically searches for configurations that are better than the current one according to an application model and explicit performance and availability targets. On finding a better configuration, an agent will enact the new configuration after a random delay to avoid possible collisions. We evaluate our framework by studying a prototype UDDI service. We show that while agents act autonomously, the service rapidly reaches a stable and appropriate configuration in response to system dynamics.

Abstract—The limitations of client/server systems become evident in large scale distributed environments. In such systems individual resources are concentrated on one or a small number of nodes and in order to provide access with acceptable response times sophisticated load balancing and fault-tolerance algorithms have to be applied. Also limitation on the network bandwidth adds the bottleneck problem. These problems have motivated researchers to come up with approaches to distribute processing loads and network bandwidth among all nodes participating in a distributed system. Peer-to-Peer systems offer an alternative to traditional client/server systems that solve bottleneck problems but need complex algorithms. This paper provides an overview on different p2p architectures and compares them with each other regarding some issues such as scalability, fault tolerance and manageability. P2P systems constitute highly dynamic networks of peers with complex topologies that create an overlay network. P2P applications need sophisticated discovery mechanisms to enable peers to find, identify and communicate with other peers. We discuss the discovery mechanisms in p2p systems which are based on the topology of the network and study the potential challenges in p2p networks such as reliability, security and adaptability.

Balancing of structured peer-to-peer graphs, including their zone sizes, has recently become an important topic of distributed hash table (DHT) research. To bring analytical understanding into the various peer-join mechanisms based on consistent hashing, we study how zone-balancing decisions made during the initial sampling of the peer space affect the resulting zone sizes and derive several asymptotic bounds for the maximum and minimum zone sizes that hold with high probability. Several of our results contradict those of prior work and shed new light on the theoretical performance limitations of consistent hashing. We use simulations to verify our models and compare the performance of the various methods using the example of recently proposed de