Checkpointing as a Service in Heterogeneous Cloud Environments

Infrastructure-as-a-Service (IaaS) cloud computing is gaining significant interest in industry and academia as an alternative platform for running scientific applications. Given the dynamic nature of IaaS clouds and the long runtime and resource utilization of such applications, an efficient checkpoint-restart mechanism becomes paramount in this context. This paper proposes a solution to the aforementioned challenge that aims at minimizing the storage space and performance overhead of checkpoint-restart. We introduce an approach that leverages virtual machine (VM) disk-image multi-snapshotting and multi-deployment inside checkpoint-restart protocols running at guest level in order to efficiently capture and potentially roll back the complete state of the application, including file system modifications. Experiments on the G5K testbed show substantial improvement for MPI applications over existing approaches, both for the case when customized checkpointing is available at application level and the case when it needs to be handled at process level.

Proceedings of the 2010 IEEE/IFIP Network Operations and Management Symposium, NOMS 2010, 2010

Crash and omission failures are common in service providers: a disk can break down or a link can fail anytime. In addition, the probability of a node failure increases with the number of nodes. Apart from reducing the provider's computation power and jeopardizing the fulfillment of his contracts, this can also lead to computation time wasting when the crash occurs before finishing the task execution. In order to avoid this problem, efficient checkpoint infrastructures are required, especially in virtualized environments where these infrastructures must deal with huge virtual machine images.

International Journal of Networking and Computing, 2019

Input/output (I/O) from various sources often contend for scarcely available bandwidth. For example, checkpoint/restart (CR) protocols can help to ensure application progress in failureprone environments. However, CR I/O alongside an application's normal, requisite I/O can increase I/O contention and might negatively impact performance. In this work, we consider different aspects (system-level scheduling policies and hardware) that optimize the overall performance of concurrently executing CR-based applications that share I/O resources. We provide a theoretical model and derive a set of necessary constraints to minimize the global waste on a given platform. Our results demonstrate that Young/Daly's optimal checkpoint interval, despite providing a sensible metric for a single, undisturbed application, is not sufficient to optimally address resource contention at scale. We show that by combining optimal checkpointing periods with contention-aware system-level I/O scheduling strategies, we can significantly improve overall application performance and maximize the platform throughput. Finally, we evaluate how specialized hardware, namely burst buffers, may help to mitigate the I/O contention problem. Overall, these results provide critical analysis and direct guidance on how to design efficient, CR ready, large-scale platforms without a large investment in the I/O subsystem.

Scalable Computing: Practice and Experience, 2014

Cloud computing is a new benchmark towards enterprise application development that can facilitate the execution of workflows in business process management system. The workflow technology can manage the business processes efficiently satisfying the requirements of modern enterprises. Besides the scheduling, the fault tolerance is a very important issue in the workflow management. In this paper, we analyse and compare between some existing checkpointing strategies, then we propose a lightweight checkpointing adequate to the cloud computing and the workflows characteristics. The proposed strategy is an Adaptive Time based Coordinated Checkpointing ATCCp, it ensures a strong consistency without any synchronization. ATCCp uses the concept of soft checkpointing to minimize the storage time and it uses the VIOLIN topology to improve the checkpointing performances. According to the experimental results, our approach decreases the overhead and the SLA violations.

2015 IEEE 40th Local Computer Networks Conference Workshops (LCN Workshops), 2015

Cloud computing services have varying performance characteristics that the cloud provider often hides from the user. Thus, it is difficult for a user to make operational decisions about when and where to run their jobs. In this paper, we present a series of experiments on a cloud computing platform to understand these characteristics and then present a series of strategies to optimize utilization on these cloud platforms. Specifically, our experiments were performed on the lowest tier of Amazon Elastic Cloud (EC2) resources, and initial tests measured performance characteristics of these resources at different locations over time. Testing revealed that certain performance measures could be improved by location choice. For short-duration jobs, CPU performance was nearly constant, but storage performance had more variability. For long-duration jobs, CPU throttling caused a significant penalty. Using a system of checkpoint and migration between virtual machines allowed this CPU penalty to be avoided, resulting in significant savings and improved runtime.

Anais do XXIV Workshop de Testes e Tolerância a Falhas (WTF 2023)

This paper concisely reviews checkpointing techniques in distributed systems, focusing on various aspects such as coordinated and uncoordinated checkpointing, incremental checkpoints, fuzzy checkpoints, adaptive checkpoint intervals, and kernel-based and user-space checkpoints. The review highlights interesting points, outlines how each checkpoint approach works, and discusses their advantages and drawbacks. It also provides a brief overview of the adoption of checkpoints in different contexts in distributed computing, including Database Management Systems (DBMS), State Machine Replication (SMR), and High-Performance Computing (HPC) environments. Additionally, the paper briefly explores the application of checkpointing strategies in modern cloud and container environments, discussing their role in live migration and application state management. The review offers valuable insights into their adoption and application across various distributed computing contexts by summarizing the hi...

Future Generation Computer Systems, 2010

In recent years there has been a significant effort to develop middleware that facilitates the execution of applications on Grid infrastructures. However, support for fault-tolerant execution continues to be scarce. The CPPC-G framework is a service-based architecture designed to provide efficient faulttolerant mechanisms for the execution of sequential and parallel applications on grids. Applications to be managed by CPPC-G are expected to be preprocessed with CPPC (ComPiler for Portable Checkpointing), a tool for automatically inserting portable checkpoint instrumentation into the code of parallel applications. Built on top of existing Globus services, CPPC-G services are in charge of submitting and monitoring CPPC applications, managing generated checkpoint files, detecting failures and automatically restarting failed executions. In this paper the feasibility of this approach is assessed by measuring the performance of CPPC-G, quantitatively addressing its impact on application performance. Results show that the increase in overall throughput and availability comes with minor performance degradation. (G. Rodríguez), pardo@udc.es (X.C. Pardo), mariam@udc.es (M.J. Martín), pglez@udc.es (P. González). range of different machines. Portability in this context means to provide the following fundamental features:

2012 Second International Conference on Cloud and Green Computing, 2012

Cloud computing offers new computing paradigms, capacity, and flexibility to high performance computing (HPC) applications with provisioning of a large number of Virtual Machines (VMs) for computation-intensive applications using the Hardware as a Service (HaaS) model. Due, however, to the large number of VMs and electronic components in HPC systems in the cloud, any fault during the execution would result in re-running the application, which will cost time, money and energy. In this paper we present a proactive Fault Tolerance (FT) approach to HPC systems in the cloud to reduce the wall clock execution time in the presence of faults. We develop a generic FT algorithm for HPC systems in the cloud. Our algorithm does not rely on a spare node prior to prediction of a failure. We analyze the dollar cost of provisioning spare nodes to assess the value of our approach. Our experimental results obtained from a real cloud execution environment show that the wall clock execution time of the computation-intensive applications in cloud can be reduced by as much as 30%. The frequency of checkpointing of computation-intensive applications can be reduced to 50% with our fault tolerance approach for HPC in the cloud, compared to current FT approaches.

Computational clusters, the grids that federate them, and the applications that utilize their significant computing potential, all continue to grow with advances in hardware technology, cluster management, and grid middleware solutions. As they do, the likelihood that large-scale long-running grid and cluster applications will have to deal with underlying node unavailability and cluster failure increases as well. The primary weapons against this problemcheckpointing, migration, replication, and effective scheduling-do not currently scale well enough to be effective for the largest, most important grid and cluster applications. Complementary research efforts in upstate New York are beginning to address this issue at a variety of levels, including: (i) low level mechanisms that will predict individual processor failures by observing and reacting to low-level indicators in their chip state; (ii) scalable cluster-level checkpointing solutions that do not require centralized storage for replicated checkpoints; (iii) grid-level efforts to differentiate between different node unavailability states, to characterize the behavior of nodes, to predict their near-future unavailability, and to make better grid scheduling decisions based on this information, and on characteristics and capabilities of applications.

2016

System-level checkpoint-restart is a critical technology for long-running jobs in high-performance computing. Yet, only two approaches to checkpointing MPI applications continue to survive in wide use today. One approach is to use the kernel module-based BLCR in combination with an MPI checkpoint-restart service particular to the MPI implementation in use. Unfortunately, this lacks support for some important Linux system services such as SysV IPC (e.g., shared memory objects). A second approach has been to use the original 2009 DMTCP implementation (herein referred to as DMTCP-09) for transparent, system-level checkpointing. Unfortunately, DMTCP-09 lacked support for checkpointing many of the necessary features found by MPI in a modern batch environment. These include: ssh, the InfiniBand network, process migration (restarting an MPI application on different cluster nodes), and modified file path prefixes on restart (typically due to a changing current directory, mount points, libra...