Showing papers by "Hans P. Reiser published in 2008"

Efficient state transfer for hypervisor-based proactive recovery

Abstract: Proactive recovery of replicated services is a novel approach that allows tolerating a potentially unlimited number of malicious faults during system lifetime by periodically restarting replicas from a correct state. Recovering a stateful replica requires a time-consuming transfer and verification of the state. During this time, the replica usually is unable to handle client requests. Our VM-FIT architecture harnesses virtualization to significantly reduce this service unavailability. Our approach allows recovery in parallel with service execution, and uses copy-on-write techniques and provides efficient state transfer support between virtual replicas on a host.

Adaptive web service migration

Abstract: In highly dynamic and heterogeneous environments such as mobile and ubiquitous computing, software must be able to adapt at runtime and react to the environment. Furthermore it should be independent of a certain hardware platform and implementation language. In this paper, we propose an infrastructure for self-adaptive migratable Web services (SAM-WS) for implementing applications for such environments. A SAM-WS supports stateful migration and adaptation to particular application context by being able to dynamically change the interface, locally available state and implementation in use. Despite adaptation and migration it maintains a unique ID during the whole life time. This allows clients to have a location-independent reference to a specific Web service instance. Although our prototype implementation is based on Apache Axis, the concept can be easily ported to any Web service framework without platform modifications. We provide an example application and performance measurements for different system platforms ranging from a standard device to resource-restricted mobile devices.

Fault and Intrusion Tolerance on the Basis of Virtual Machines

Abstract: Fault and intrusion tolerance is an important paradigm for building distributed systems that work in spite of accidental and malicious faults. This paper discusses how to harness virtualization technology for building such dependable systems. We show that virtualization promotes a hybrid fault model that allows tolerating malicious intrusions in application domains with little overhead. The proposed architecture features mechanisms for supporting heterogeneity of the replicas. A hypervisor-based replication controller achieves perpetual operation through periodic proactive recovery of the replicas. Re-mapping of state storage between virtual machines speeds up the state transfer of a stateful replicated service. Our VM-FIT prototype implements the core functionality of such a virtualization-based replication architecture. We present some performance measurements and close with a discussion of future research directions.

Multithreading strategies for replicated objects

Abstract: Replicating objects usually requires deterministic behaviour for maintaining a consistent state. Multithreading is a critical source of non-determinism, completely unsupported in most fault-tolerant middleware systems. Recent publications have defined deterministic scheduling algorithms that operate at the middleware level and allow multithreading for replicated objects. This approach avoids deadlocks, improves performance, and makes the development better resemble that of non-replicated objects. This paper surveys those algorithms and analyses their differences. It also defines extensions to two efficient multithreading algorithms to support nested invocations and condition variables with time-bounded wait operations similar to the Java synchronisation model. In addition, we provide an experimental evaluation and performance comparison of the algorithms, indicating the areas in which each algorithm performs best. We conclude that replication middleware should implement reconfigurable multithreading strategies, as there is no optimal one-size-fits-all solution.

FOREVER: Fault/intrusiOn REmoVal through Evolution & Recovery

Abstract: The goal of the FOREVER project is to develop a service for Fault/intrusiOn REmoVal through Evolution & Recovery. In order to achieve this goal, our work addresses three main tasks: the definition of the FOREVER service architecture; the analysis of how diversity techniques can improve resilience; and the evaluation of the FOREVER service. The FOREVER service is an important contribution to intrustion-tolerant replication middleware and significantly enhances the resilience.

Using object replication for building a dependable version control system

Abstract: Object-oriented technologies are frequently used to design and implement distributed applications. Object replication is a well-established approach to increase the dependability for such applications. Generic replication infrastructures often fail to meet non-standard application-specific requirements such as support for client-side computing. Our FTflex replication infrastructure combines the fragmented object model with semantic annotations in order to customize and optimize replication mechanisms, and thus provides a more flexible replication infrastructure. This paper presents DiGit, a replicated version control system based on the architecture of Git. DiGit is implemented with the help of the FTflex infrastructure for object replication. The contributions of this paper are twofold. First, the paper evaluates the fitness of our replication framework for a specific, complex application. We identify two advantages of the replication infrastructure: the ability to provide client-side code as a conceptually integral part of a remote service, and support for an optimized protocol for remote interaction. As a second contribution, the paper presents a powerful replicated version control system and shows the lessons learned from using object replication in such a system.