Luiz A. Rodrigues

A Robust Permission-Based Hierarchical Distributed k-Mutual Exclusion Algorithm

Distributed mutual exclusion is a basic building block of distributed systems that coordinates the access to critical shared resources. This work introduces a novel permission-based k-mutual exclusion algorithm for distributed systems with crash faults. Processes monitor each other and organize themselves on an adaptive virtual topology that is based on the hypercube and presents several logarithmic properties. Mutual exclusion is deployed on top of this monitoring system. Processes communicate through spanning trees which are created with a fully distributed strategy that tolerates faults by using process state information provided by the underlying monitoring system. Both the mutual exclusion and the distributed spanning tree algorithm are formally specified. The strategy is proven to guarantee the safety and liveness of the concurrent access of n processes to k critical resources. Experimental results are presented, showing that the algorithm performs efficiently even when up to n-1 processes are faulty.

A Publish/Subscribe System Using Causal Broadcast over Dynamically Built Spanning Trees

In this paper we present VCube-PS, a topic-based Publish/Subscribe system built on the top of a virtual hypercubelike topology. Membership information and published messages to subscribers (members) of a topic group are broadcast over dynamically built spanning trees rooted at the message’s source. For a given topic, delivery of published messages respects causal order. Performance results of experiments conducted on the PeerSim simulator confirm the efficiency of VCube-PS in terms of scalability, latency, number, and size of messages when compared to a single rooted, not dynamically, tree built approach.

An Autonomic Implementation of Reliable Broadcast Based on Dynamic Spanning Trees

Reliable broadcast is a basic building block of dependable distributed systems that allows the dissemination of messages to all processes guaranteeing that either none or all correct processes deliver the message, despite the occurrence of failures. In this work we introduce an autonomic strategy to implement reliable broadcast. Processes self-organize themselves on an overlay based on a hypercube-like topology. The overlay is maintained by the execution of an underlying monitoring algorithm that guarantees several logarithmic properties even in the presence of processes failures. The reliable broadcast strategy employs spanning trees that are dynamically built embedded in the overlay. Nodes fail by crashing, crashes are permanent, and the broadcast strategy works correctly for an arbitrary number of process crashes. After a process crashes, the reconstruction of the spanning trees is transparent, not inducing, therefore, performance degradation. Besides the formal specification, we also present experimental results obtained with simulation.

A Communication-Efficient Causal Broadcast Protocol

A causal broadcast ensures that messages are delivered to all nodes (processes) preserving causal relation of the messages. In this paper, we propose a causal broadcast protocol for distributed systems whose nodes are logically organized in a virtual hypercube-like topology called VCube. Messages are broadcast by dynamically building spanning trees rooted in the messages source node. By using multiple trees, the contention bottleneck problem of a single root spanning tree approach is avoided. Furthermore, different trees can intersect at some node. Hence, by taking advantage of both the out-of-order reception of causally related messages at a node and these paths intersections, a node can delay to one or more of its children in the tree, the forwarding of the messages whose some causal dependencies it knows that the children in question can not satisfy yet. Such a delay does not induce any overhead. Experimental evaluation conducted on top of PeerSim simulator confirms the communication effectiveness of our causal broadcast protocol in terms of latency and message traffic reduction.

A distributed k-mutual exclusion algorithm based on autonomic spanning trees

Abstract Distributed k -mutual exclusion ensures that at most a single process has permission to access each of the k copies of a critical resource. In this work we present an autonomic solution for distributed k -mutual exclusion that adapts itself after system changes. Our solution employs a hierarchical best-effort broadcast algorithm to propagate messages reliably and efficiently. The broadcast is based on another autonomic building block: a distributed algorithm for creating and maintaining spanning trees constructed in a fully distributed and adaptive way on top of a virtual hypercube-like topology, called VCube. The proposed solutions are autonomic in the sense that they reconfigure themselves automatically after the detection of faults given the set of correct processes in the system. All proposed algorithms are described, specified, and proofs of correctness are given. Results from simulation show that the proposed approach is more efficient and scalable compared to other solutions.

Parallel cut tree algorithms

Abstract A cut tree is a combinatorial structure that represents the edge-connectivity between all pairs of vertices of an undirected graph. Cut trees solve the all pairs minimum s – t -cut problem efficiently. Cut trees have a large number of applications including the solution of important combinatorial problems in fields such as graph clustering and graph connectivity. They have also been applied to scheduling problems, social network analysis, biological data analysis, among others. Two sequential algorithms to compute a cut tree of a capacitated undirected graph are well known: the Gomory–Hu algorithm and the Gusfield algorithm. In this work three parallel cut tree algorithms are presented, including parallel versions of Gusfield and Gomory–Hu algorithms. A hybrid algorithm that combines techniques from both algorithms is proposed which provides a more robust performance for arbitrary instances. Experimental results show that the three algorithms achieve significant speedups on real and synthetic graphs. We discuss the trade-offs between the alternatives, each of which presents better results given the characteristics of the input graph. On several instances the hybrid algorithm outperformed both other algorithms, being faster than the parallel Gomory–Hu algorithm on most instances.

An Autonomic Hierarchical Reliable Broadcast Protocol for Asynchronous Distributed Systems with Failure Detector

Reliable broadcast is a fundamental building block in fault-tolerant distributed systems. It consists of a basic primitive that provides agreement among processes of the system on the delivery of each broadcast message, i.e., either none or all correct processes deliver the message, despite failures of processes. In this work, we propose a reliable broadcast solution on top of VCube, assuming that the system is asynchronous. VCube is an autonomic monitoring layer that organizes processes on a hypercube-like overlay which provides several logarithmic properties even in the presence of processes failures. We consider that processes can fail by crashing, do not recover, and faults are eventually detected by all correct processes. The protocol tolerates false suspicions by sending additional messages to suspected processes but logarithmic properties of the algorithm are still kept. Experimental results show the efficiency of the proposed solution compared to an one-to-all strategy.

An Autonomic Majority Quorum System

A quorum system is a collection of process sets (quorums) which intersect among themselves. Quorums are used by many distributed applications such as mutual exclusion, data replication, and for the dissemination of information. This work presents an autonomic solution to build a majority quorum system that is self-adaptive by dymically reconfiguring itself if processes fail. Processes can fail by crashing and crashes are permanent. The proposed solution is defined over a VCube - a virtual hypercube-like topology [1], and uses failure information from the VCubes monitoring system. Each quorum is built so that any quorum has a majority of the faulty-free processes of the system. Upon the detection of a failure, the quorum system reconfigures automatically, tolerating up to n-1 crashed processes. Experimental results confirm the efficiency of the proposed algorithm compared with other solutions of the literature.

Distributed and Autonomic Minimum Spanning Trees

This paper presents an autonomic algorithm for constructing minimum spanning trees in a distributed system. Processes are virtually organized on a hypercube-like topology, called VCube. The spanning trees are dynamically created from any source process and regenerated transparently when processes become faulty. Any tree reconstruction is autonomously done based on the virtual topology and on fault information, supporting up to n-1 faulty processes. Two applications using the autonomic spanning tree algorithm are proposed, one for the best-effort broadcast and another for reliable broadcast. Besides the formal specification, simulation results are presented, including comparison results with other alternative solutions.

A Parallel Implementation of Gomory-Hu's Cut Tree Algorithm

Cut trees are a compact representation of the edge-connectivity between every pair of vertices of an undirected graph, and have a large number of applications. In this work a parallel version of the well known Gomory-Hu cut tree algorithm is presented. The parallel strategy is based on the master/slave model. The strategy is optimistic in the sense that the master process manipulates the tree being constructed and the slaves solve minimum s-t-cuts independently. Another version is proposed that employs a heuristic that enumerates all (up to a limit) of the minimum s-t-cuts in order to choose the most balanced one. The algorithm was implemented and extensive experimental results are presented, including a comparison with Gusfieldâs cut tree algorithm. Parallel versions of these algorithms have achieved significant speedups on real and synthetic graphs. We discuss the trade-offs between the two alternatives, each of which presents better results given the characteristics of the input graph. In particular, the existence of balanced cuts clearly gives an advantage to Gomory-Huâsalgorithm.