this paper, we use the terms event logging and message logging interchangeably

Diskless Checkpointing is a technique for checkpointing the state of a long-running computation on a distributed system without relying on stable storage. As such, it eliminates the performance bottleneck of traditional checkpointing on distributed systems. In this paper, we motivate diskless checkpointing and present the basic diskless checkpointing scheme along with several variants for improved performance. The performance of the basic scheme and its variants is evaluated on a high-performance network of workstations and compared to traditional disk-based checkpointing. We conclude that diskless checkpointing is a desirable alternative to disk-based checkpointing that can improve the performance of distributed applications in the face of failures.

Checkpointing is a useful technique for rollback recovery of parallel applications. While extensive research has been performed on checkpointing in parallel environments, there are few checkpointers available to application users on commercial parallel computers. This paper presents one such checkpointer: CLIP. CLIP is a user-level library that provides semitransparent checkpointing for parallel programs on the Intel Paragon multicomputer. It is publicly available to Paragon users at no cost. Conceptually, checkpointing a multicomputer is quite straightforward. However, when creating an actual tool for checkpointing a complex machine like the Paragon, many more issues arise that require careful design decisions to be made. Sometimes ease-of-use must be sacrificed for efficiency and/or correctness. This paper details what these decisions are, and how they were made in CLIP. We also present performance data when checkpointing several long-running Paragon applications with CLIP. The bottom line is that a convenient, general-purpose checkpointing tool like CLIP can provide fault-tolerance on a massively parallel multicomputer like the Paragon with very good performance.

We discuss the design and implementation of Egida, an objectoriented toolkit designed to support transparent rollback-recovery. Egida exports a simple specification language that can be used to express arbitrary rollback recovery protocols. From this specification, Egida automatically synthesizes an implementation of the specified protocol by gluing together the appropriate objects from an available library of “building blocks”. Egida is extensible and facilitates rapid implementation of rollback recovery protocols with minimal programming effort. We have integrated Egida with the MPICH implementation of the MPI standard. Existing MPI applications can take advantage of Egida without any modifications: fault-tolerance is achieved transparently—all that is needed is a simple re-link of the MPI application with Egida. 1

Utilizing a collection of workstations and supercomputers in a metacomputing environment does not only offer an enormous amount of computing power, but also raises new problems. The true potential of WAN-based distributed computing can only be exploited if the application-to-architecture mapping reflects the different processor speeds, network performances and the application's communication characteristics. In this paper, we present the Metacomputer Adaptive Runtime System MARS,a framework for minimizing the execution time of distributed applications on a WAN metacomputer. Work-load balancing and task migration is based on dynamic information on the processor load and network performance. Moreover, MARS uses accumulated statistical data on previous execution runs of the same application to derive an improved task-to-process mapping. Migration decisions are based on (1) the current system load, (2) the network load and (3) previously obtained application-specific characteri...

Coordinated checkpointing systems are popular and general-purpose tools for implementing process migration, coarse-grained job swapping, and fault-tolerance on networks of workstations. Though simple in concept, there are several design decisions concerning the placement of checkpoint files that can impact the performance and functionality of coordinated checkpointers. Although several such checkpointers have been implemented for popular programming platforms like PVM and MPI, none have taken this issue into consideration. This paper addresses the issue of checkpoint placement and its impact on the performance and functionality of coordinated checkpointing systems. Several strategies, both old and new, are described and implemented on a network of SPARC-5 workstations running PVM. These strategies range from very simple to more complex, borrowing heavily from ideas in RAID (Redundant Arrays of Inexpensive Disks) faulttolerance. The results of this paper will serve as a guide so that f...

Performance prediction of checkpointing systems in the presence of failures is a well-studied research area. While the literature abounds with performance models of checkpointing systems, none address the issue of selecting runtime parameters other than the optimal checkpointing interval. In particular, the issue of processor allocation is typically ignored. In this paper, we present a performance model for long-running parallel computations that execute with checkpointing enabled. We then discuss how it is relevant to today's parallel computing environments and software, and present case studies of using the model to select runtime parameters. Keywords: Checkpointing, performance prediction, parameter selection, parallel computation, Markov chain, exponential failure and repair distributions. 1

In this paper we discuss a new approach to dynamic load balancing of parallel jobs in clusters of workstations and describe the implementation into a Unix run-time environment. The efficiency of the proposed methodology is shown by means of a number of case studies. Key words: Dynamic load balancing, task migration, distributed computing systems, clusters of workstations. 1 Introduction With the advent of high-speed networks (most prominently FDDI and ATM), clusters of workstations achieve the same scalable parallelism as the current MPP architectures. However, software support for such parallel cluster systems still lags behind. These parallel cluster systems require new programming paradigms and environments to provide the user with mechanisms and tools to exploit the full potential of the available distributed resources. In these systems changes such as variation in demand of processor power, variation in number of processors, or dynamic changes in the run-time behaviour of ...

Networks of workstations (NOWs) offer a cost effective platform for high-performance, long-running parallel computations. However, these computations must be able to tolerate the changing and often faulty nature of NOW environments. We present high-performance implementations of several fault-tolerant algorithms for distributed scientific computing. The fault-tolerance is based on diskless checkpointing, a paradigm that uses processor redundancy rather than stable storage as the fault-tolerant medium. These algorithms are able to run on clusters of workstations that change over time due to failure, load or availability. As long as there are at least n processors in the cluster, and failures occur singly, the computation will complete in an efficient manner. We discuss the details of how the algorithms are tuned for fault-tolerance and present the performance results on a PVM network of Sun workstations connected by a fast, switched ethernet.

This paper describes the design, implementation, and evaluation of a run-time system for clusters of workstations that allows the rapid testing of checkpoint protocols with standard benchmarks. To achieve this goal, RENEW provides a flexible set of operations that facilitates the integration of a protocol in the system with reduced programming effort. To support a broad range of applications, RE-NEW exports, as its external interface, the industry endorsed Message Passing Interface (MPI). Three distinct classes of protocols were evaluated using the RENEW environment with SPEC and NAS benchmarks on a network of workstations connected by ATM. It was observed that the communication-induced protocol emulated the behavior of the coordinated protocol, with comparable performance. The message logging protocol degraded the performance. Even though the message logging protocol was slower due to log replay, all three protocols required a similar amount of time to restore the applicationto the same state as before failure occurred and recovery was initiated. 1