Abstract not found

An e-Transaction is one that executes exactly-once despite failures. This paper describes a distributed protocol that implements the abstraction of e-Transactions in three-tier architectures. Three-tier architectures are typically Internet-oriented architectures, where the end-user interacts with front-end clients (e.g., browsers) that invoke middle-tier application servers (e.g., web servers) to access back-end databases. We implement the e-Transaction abstraction using an asynchronous replication scheme that preserves the three-tier nature of the architecture and introduces a very acceptable overhead with respect to unreliable solutions. 1

Total order multicast greatly simplifies the implementation of fault-tolerant services using the replicated state machine approach. The additional latency of total ordering can be masked by taking advantage of spontaneous ordering observed in LANs: A tentative delivery allows the application to proceed in parallel with the ordering protocol. The effectiveness of the technique rests on the optimistic assumption that a large share of correctly ordered tentative deliveries offsets the cost of undoing the effect of mistakes. This paper proposes a simple technique which enables the usage of optimistic delivery also in WANs with much larger transmission delays where the optimistic assumption does not normally hold. Our proposal exploits local clocks and the stability of network delays to reduce the mistakes in the ordering of tentative deliveries. An experimental evaluation of a modified sequencer-based protocol is presented, illustrating the usefulness of the approach in fault-tolerant database management.

This paper presents the semi-passive replication technique -- a variant of passive replication -- that can be implemented in the asynchronous system model without requiring a membership service to agree on a primary. Passive

The Paxos part-time parliament protocol of Lamport provides a very practical way to implement a fault-tolerant deterministic service by replicating it over a distributed message passing system. The contribution of this paper is a faithful deconstruction of Paxos that preserves its efficiency in terms of forced logs, messages and communication steps. The key to our faithful deconstruction is the factorisation of the fundamental algorithmic principles of Paxos within two abstractions: weak leader election and round-based consensus, itself based on a round-based register abstraction. Using those abstractions, we show how to reconstruct, in a modular manner, known and new variants of Paxos. In particular, we show how to (1) alleviate the need for forced logs if some processes remain up for sufficiently long, (2) augment the resilience of the algorithm against unstable processes, (3) enable single process decision with shared commodity disks, and (4) reduce the number of communication steps during stable periods of the system.

In this paper, we propose a new architecture for group communication middleware. Current group communication systems share some common features, despite the big di#erences that exist among them. We first point out these common features by describing the most representative group communication architectures implemented over the last 15 years. Then we show the features of our new architecture, which provide several advantages over the existing architectures: (1) it is less complex, (2) it defines a set of group communication abstractions that is more consistent than the abstractions usually provided, and (3) it can be made more responsive in case of failures.

In this work we present a novel protocol for total ordering of messages in asynchronous distributed environments prone to machine and communication link failures. Using the protocol as a building block, we constructed a Totally Ordered Group Communication (TOGC) system, i.e., a group communication service with a totally ordered multicast primitive. TOGC is a powerful infrastructure for building distributed fault-tolerant applications such as totally ordered broadcast, consistent object replication, distributed shared memory, Computer Supported Cooperative Work (CSCW) applications and distributed monitoring and display applications. An important contribution of the total ordering protocol described in this work is its ability to dynamically adjust the message delivery flow to changes in the transmission rates of the participating processes. The adaptation is accomplished by assigning delivery priorities (weights) to messages according to sender transmission rates. The priorities are det...

Providing reliable group communication is an ever recurring topic in distributed settings. In mobile ad hoc networks, this problem is even more significant since all nodes act as peers, while it becomes more challenging due to highly dynamic and unpredictable topology changes. In order to overcome these difficulties, we deviate from the conventional point of view, i.e., we "fight fire with fire," by exploiting the nondeterministic nature of ad hoc networks. Inspired by the principles of gossip mechanisms and probabilistic quorum systems, we present in this paper PILOT (ProbabilistIc Lightweight grOup communication sysTem) for ad hoc networks, a two-layer system consisting of a set of protocols for reliable multicasting and data sharing in mobile ad hoc networks. The performance of PILOT is predictable and controllable in terms of both reliability (fault tolerance) and efficiency (overhead). We present an analysis of PILOT's performance, which is used to fine-tune protocol parameters to obtain the desired trade off between reliability and efficiency. We confirm the predictability and tunability of PILOT through simulations with ns-2.

The paper is a tutorial on fault-tolerance by replication in distributed systems. We start by defining linearizability as the correctness criterion for replicated services (or objects), and present the two main classes of replication techniques: primary-backup replication and active replication. We introduce group communication as the infrastructure providing the adequate multicast primitives to implement either primary-backup replication, or active replication. Finally, we discuss the implementation of the two most fundamental group multicast primitives: total order multicast and view synchronous multicast.

A replication logic is the set of protocols and mechanisms implementing a software replication technique. A three-tier approach to replication consists in separating the replication logic from both clients and replicated servers by embedding such logic in a middle-tier.