Rida A. Bazzi

The power of processor consistency

Shared memories that provide weaker consistency guarantees than the traditional sequentially consistent or atomic memories have been claimed to provide the key to building scalable systems. One influential memory model, processor considency, has been cited widely in the literature but, due to the lack of a precise and formal definition, contradictory claims have been made regarding its power. We use a formal model to give two distinct definitions of processors consistency: one corresponding to Goodman’s original proposal and the other corresponding that given by the implementors of the DASH system. These definitions are non-operational and can be easily related to other types of memories. To illustrate the power of processor consistency, we exhibit a non-cooperative solution to the mutual exclusion problem that is correct with processor consistency. As a contrast, we show that Lamport’s Bakery algorithm is not correct with processor consistency.

On the establishment of distinct identities in overlay networks

We study ways to restrict or prevent the damage that can be caused in a peer-to-peer network by corrupt entities creating multiple pseudonyms. We show that it is possible to remotely issue certificates that can be used to test the distinctness of identities. To our knowledge, this is the first work that shows that remote anonymous certification of identity is possible under adversarial conditions. Our certification protocols are based on geometric techniques that establish location information in a fault-tolerant and distributed fashion. They do not rely on a centralized certifying authority or infrastructure that has direct knowledge of entities in the system, and work in Euclidean or spherical geometry of arbitrary dimension. Our protocols tolerate corrupt entities, including corrupt certifiers as well as collusion by certification applicants and certifiers. We consider both broadcast and point-to-point message passing models.

Synchronous Byzantine quorum systems

Summary. Quorum systems have been used to implement many coordination problems in distributed systems such as mutual exclusion, data replication, distributed consensus, and commit protocols. Malkhi and Reiter recently proposed quorum systems that can tolerate Byzantine failures; they called these systems Byzantine quorum systems and gave some examples of such quorum systems. In this paper, we propose a new definition of Byzantine quorums that is appropriate for synchronous systems. We show how these quorums can be used for data replication and propose a general construction of synchronous Byzantine quorums using standard quorum systems. We prove tight lower bounds on the load of synchronous Byzantine quorums for various patterns of failures and we present synchronous Byzantine quorums that have optimal loads that match the lower bounds for two failure patterns.

Non-skipping timestamps for byzantine data storage systems

We study the problem of implementing a replicated data store with atomic semantics for non self-verifying data in a system of n servers that are subject to Byzantine failures. We present a solution that significantly improves over previously proposed solutions. Timestamps used by our solution cannot be forced to grow arbitrarily large by faulty servers as is the case for other solutions. Instead, timestamps grow no faster than logarithmically in the number of operations. We achieve this saving by defining and providing an implementation for non-skipping timestamps, which are guaranteed not to skip any value. Non-skipping timestamps allow us to reduce the space requirements for readers to O(max|Q|), Where |Q| ≤ n. This is a significant improvement over the best previously known solution which requires O(fn) space, where f is the maximum number of faulty servers in the system. The solution we present has a low write-load if f is small compared to n, whereas the previously proposed solution always has a high constant write-load.

On the Availability of Non-strict Quorum Systems

Allowing read operations to return stale data with low probability has been proposed as a means to increase availability in quorums systems. Existing solutions that allow stale reads cannot tolerate an adversarial scheduler that can maliciously delay messages between servers and clients in the system and for such a scheduler existing solutions cannot enforce a bound on the staleness of data read. This paper considers the possibility of increasing system availability while at the same time tolerating a malicious scheduler and guaranteeing an upper bound on the staleness of data. We characterize the conditions under which this increase is possible and show that it depends on the ratio of the write frequency to the servers’ failure frequency. For environments with a relatively large failure frequency compared to write frequency, we propose K-quorums that can provide higher availability than the strict quorum systems and also guarantee bounded staleness. We also propose a definition of k-atomicity and present a protocol to implement a k-atomic register using k-quorums.

Bounded wait-free implementation of optimally resilient byzantine storage without (unproven) cryptographic assumptions

We present the first optimally resilient, bounded, wait-free implementation of a distributed atomic register, tolerating Byzantine readers and (up to one-third of) Byzantine servers, without the use of unproven cryptographic primitives or requiring communication among servers. Unlike previous (non-optimal) solutions, the sizes of messages sent to writers depend only on the actual number of active readers and not on the total number of readers in the system. With a novel use of secret sharing techniques combined with write back throttling we present the first solution to tolerate Byzantine readers information theoretically, without the use of cryptographic techniques based on unproven number-theoretic assumptions.

Simplifying fault-tolerance: providing the abstraction of crash failures

The difficulty of designing fault-tolerant distributed algorithms incr eases with the severity of failures that an algorithm must tolerate, especially for systems with synchronous message passing. This paper considers methods that automatically translate algorithms tolerant of simple crash failures into ones tolerant of more severe failures. These translations simplify the design task by allowing algorithm designers to assume that processors fail only by stopping. Such translations can be quantified by two measures: fault-tolerance, which is a measure of how many processors must remain correct for the translation to be correct, and round-complexity, which is a measure of how the translation increases the running time of an algorithm. Understanding these translations and their limitations with respect to these measures can provide insight into the relative impact of different models of faculty behavior on the ability to provide fault-tolerant applications for systems with synchronous message passing. This paper considers translations fr om crash failures to each of the following types of more severe failures: omission to send messages; omission to send and receive messages; and totally arbitrary behavior. It shows that previously developed translaions to send-omission failures are optimal with respect to both fault-tolerance and round-complexity. It exhibits a hierarchy of translations to general (send/receive) omission failures that improves upon the fault-tolerance of previously developed translations. These translations are optimal in that they cannot be improved with respect to one measure without negatively affecting the other; that is, the hierarchy of translations is matched by corresponding hierarchy of impossibility results. The paper also gives a hierarchy of translations to arbitrary failures that improves upon the round-complexity of previously developed translations. These translations are near-optimal;

Using knowledge to optimally achieve coordination in distributed systems

Abstract A distributed computing system consists of a set of individual processors that communicate through some medium. Coordinating the actions of such processors is essential in distributed computing. Researchers have long endeavored to find efficient solutions to a variety of coordination problems. Recently, processor knowledge has been used to characterize such solutions and to derive more efficient ones. Most of this work has concentrated on the relationship between common knowledge and simultaneous coordination. This paper considers non-simultaneous coordination problems. The results of this paper add to our understanding of the relationship between knowledge and the different requirements of coordination problems. This paper considers the ideas of optimal and optimum solutions to a coordination problem and precisely characterizes the problems for which optimum solutions exist. This characterization is based on combinations of eventual common knowledge and continual common knowledge . The paper then considers more general problems, for which optimal, but no optimum, solutions exist. It defines a new form of knowledge, called extended knowledge , which combines eventual and continual knowledge, and shows how extended knowledge can be used to both characterize and construct optimal protocols for coordination.

Planar quorums

Quorum systems are used to implement many coordination problems in distributed systems such as mutual exclusion, data replication, distributed consensus, and commit protocols. This paper presents a new class of quorum systems based on connected regions in planar graphs. This class has an intuitive geometric nature and is easy to visualize and map to the system topology. We show that for triangulated graphs, the resulting quorum systems are non-dominated, which is a desirable property. We study the performance of these systems in terms of their availability, load, and cost of failures. We formally introduce the concept of cost of failures and argue that it is needed to analyze the message complexity of quorum-based protocols. We show that quorums of triangulated graphs with bounded degree have optimal cost of failures. We study a particular member of this class, the triangle lattice. The triangle lattice has small quorum size, optimal load for its size, high availability, and optimal cost of failures. Its parameters are not matched by any other proposed system in the literature. We use percolation theory to study the availability of this system. c

Heterogeneous checkpointing for multithreaded applications

We present the first heterogeneous checkpointing scheme for applications using POSIX threads. The scheme relies on source code instrumentation to achieve heterogeneity. It supports various types of synchronization primitives, such as locks, semaphores, condition variables, and join operations. Unlike other non-heterogeneous checkpointing schemes proposed in the literature, our scheme supports both kernel-level and application-level threads executing as part of the same application under various scheduling policies. Also, unlike other non-heterogeneous checkpointing mechanisms proposed in the literature, our solution does not interfere with the semantics of the application and does not use signals. Test results on various hardware platforms running Solaris, Linux, and Windows NT show that the overhead of our scheme is low.