Replication Policy Research Articles

Operationalizing the Replication Standard

In response to widespread concerns about the integrity of research published in scholarly journals, several initiatives have emerged that are promoting research transparency through access to data underlying published scientific findings. Journal editors, in particular, have made a commitment to research transparency by issuing data policies that require authors to submit their data, code, and documentation to data repositories to allow for public access to the data. In the case of the American Journal of Political Science (AJPS) Data Replication Policy, the data also must undergo an independent verification process in which materials are reviewed for quality as a condition of final manuscript publication and acceptance.
 Aware of the specialized expertise of the data archives, AJPS called upon the Odum Institute Data Archive to provide a data review service that performs data curation and verification of replication datasets. This article presents a case study of the collaboration between AJPS and the Odum Institute Data Archive to develop a workflow that bridges manuscript publication and data review processes. The case study describes the challenges and the successes of the workflow integration, and offers lessons learned that may be applied by other data archives that are considering expanding their services to include data curation and verification services to support reproducible research.

SnapFiner: A Page-Aware Snapshot System for Virtual Machines

Virtual machine (VM) snapshot, enabling a VM to be resumed from a previously recorded state, is an essential part of cloud infrastructures. Unfortunately, the snapshot data are likely to be lost due to the high rate of disk failures, so that the associated VM fails to recover properly. To enhance data availability without compromising application performance upon rollback recovery, it is desired to place multiple replicas of snapshot across disperse disks. However, due to the large size of replica, it induces non-trivial storage cost when managing massive snapshots in clouds. In this paper, we investigate this problem and find out that the semantic gap existed between snapshot creation and snapshot storing is one key factor inducing high storage cost. To this end, we propose SnapFiner, a page-aware snapshot system for creating and storing massive snapshot files efficiently. First, SnapFiner acquires a fine-grained page categorization with an in-depth page exploration from three orthogonal views, thereby discovering more pages that can be excluded from the snapshot. Second, SnapFiner varies the number of replicas for different page categories based on a page-aware replication policy, achieving low storage cost without compromising availability and performance. Third, SnapFiner handles the loss of pages either intentionally dropped upon snapshot creation or unexpectedly damaged due to disk failures, enabling proper system execution after rollback recovery. We have implemented SnapFiner on QEMU/KVM to justify its practicality for Linux guests. The experimental results demonstrate that SnapFiner reduces the storage cost by 33 and 69.5 percent respectively compared to our previous work PARS and the naive approach on QEMU/KVM and HDFS.

Optimizing dynamic survivability and security of replicated data in cloud systems under co-residence attacks

Utilizing the virtualization technology, multiple virtual machines (VMs) can be created on a single physical server for different tasks, enabling cost-effective resource sharing in cloud computing systems. However, this co-resident VM architecture can be exploited by malicious attackers, posing unique survivability and security risks for cloud users. This paper addresses one of such risks called co-residence attacks, where a malicious attacker can steal or corrupt a user's sensitive information through co-residing the attacker's VM with the target user's VM on the same physical server. We model users’ data protection policy in which sensitive data are replicated and stored on different VMs to enhance data survivability. Both user's and attacker's VMs are distributed among cloud servers at random. The arrival of attacker's requests for creating VMs is modeled by a Poisson stochastic process. We propose a probabilistic model to obtain dynamic data survivability and security indices. Based on the suggested evaluation model, dynamic data replication policies are analyzed and optimized. Numerical examples are presented to demonstrate impacts of different model parameters on the dynamic data survivability and security.

"Reproducible" Research in Mathematical Sciences Requires Changes in our Peer Review Culture and Modernization of our Current Publication Approach.

The nature of scientific research in mathematical and computational biology allows editors and reviewers to evaluate the findings of a scientific paper. Replication of a research study should be the minimum standard for judging its scientific claims and considering it for publication. This requires changes in the current peer review practice and a strict adoption of a replication policy similar to those adopted in experimental fields such as organic synthesis. In the future, the culture of replication can be easily adopted by publishing papers through dynamic computational notebooks combining formatted text, equations, computer algebra and computer code.

Data Access, Transparency, and Replication: New Insights from the Political Behavior Literature

ABSTRACTDo researchers share their quantitative data and are the quantitative results that are published in political science journals replicable? We attempt to answer these questions by analyzing all articles published in the 2015 issues of three political behaviorist journals (i.e., Electoral Studies, Party Politics, and Journal of Elections, Public Opinion & Parties)—all of which did not have a binding data-sharing and replication policy as of 2015. We found that authors are still reluctant to share their data; only slightly more than half of the authors in these journals do so. For those who share their data, we mainly confirmed the initial results reported in the respective articles in roughly 70% of the times. Only roughly 5% of the articles yielded significantly different results from those reported in the publication. However, we also found that roughly 25% of the articles organized the data and/or code so poorly that replication was impossible.

A hybrid data replication strategy with fuzzy-based deletion for heterogeneous cloud data centers

At present, huge cloud-based applications have put forward higher requests for data center storage. In a large-scale Cloud environment, data replication provides an appropriate solution for managing data files, which improves data reliability and availability. In this paper, we propose a data replication algorithm called hybrid replication strategy (HRS) that is applied into replica placement, selection, and replacement steps. HRS has three main phases and is suitable for replicating data files in cloud. In the first phase, it selects the best site (i.e., that is the most central site with high number of access) for storing new replica to reduce access time. In the second phase, HRS considers the best replica node for users based on different parameters such as CPU process capability, network transmission capability, I/O capability of disks, load, and network latency. In the third phase, the replacement decision is made in order to provide better response time. HRS can ascertain the importance of valuable replicas on the basis of a fuzzy inference system with three input parameters (i.e., number of accesses, cost, and the last time the replica was accessed). The new replication policy is simulated using the CloudSim toolkit package. Our proposed mechanism replicates the data over the cloud nodes reasonably well and is easily implementable in a real environment. Experiment results prove that HRS can significantly enhance availability, performance and load balance for data-intensive applications. In addition, it stands good without increasing additional overheads.

Performance Evaluation of Replication Policies in Microservice Based Architectures

Nowadays applications tend to be executed on distributed environments provisioned using on-demand infrastructures. The use of techniques such as application containers simplifies the orchestration of complex systems. In this context, microservices based architectures offer a promising solution for what concerns software development and scalability. In this paper, we propose an approach to study the automatic scalability of microservices architectures deployed in public and private clouds. A Fluid Petri Net model describes the characterise of the platform, and a real trace drives the approach to consider a realistic scenario. Our focus is on evaluating the performances, costs and energy consumptions from both the service provider and infrastructure provider point of view.

Synergy via Redundancy

The maximum possible throughput (rate of task completion) of a multi-server system is typically the sum of the service rates of individual servers. Recent works show that task replication can boost the throughput, in particular if the service time has high variability (Cv > 1). Thus, redundancy can be used to create synergy among servers such that their overall throughput is greater than sum of individual servers. This paper seeks to find the fundamental limit of this capacity boost achieved by task replication. The optimal adaptive replication policy can be found using a Markov Decision Process (MDP) framework, but the MDP is hard to solve in general. We propose two replication policies, MaxRate and AdaRep that gradually add replicas only when needed. To quantify the optimality gap of these policies, we also derive an a upper bound on the service capacity for the two-server case.

Efficient Redundancy Techniques in Cloud and Desktop Grid Systems using MAP/G/c-type Queues

Abstract Cloud computing is continuing to prove its flexibility and versatility in helping industries and businesses as well as academia as a way of providing needed computing capacity. As an important alternative to cloud computing, desktop grids allow to utilize the idle computer resources of an enterprise/community by means of distributed computing system, providing a more secure and controllable environment with lower operational expenses. Further, both cloud computing and desktop grids are meant to optimize limited resources and at the same time to decrease the expected latency for users. The crucial parameter for optimization both in cloud computing and in desktop grids is the level of redundancy (replication) for service requests/workunits. In this paper we study the optimal replication policies by considering three variations of Fork-Join systems in the context of a multi-server queueing system with a versatile point process for the arrivals. For services we consider phase type distributions as well as shifted exponential and Weibull. We use both analytical and simulation approach in our analysis and report some interesting qualitative results.

Task Replication for Deadline-Constrained Vehicular Cloud Computing: Optimal Policy, Performance Analysis, and Implications on Road Traffic

In vehicular cloud computing (VCC) systems, the computational resources of moving vehicles are exploited and managed by infrastructures, e.g., roadside units, to provide computational services. The offloading of computational tasks and collection of results rely on successful transmissions between vehicles and infrastructures during encounters. In this paper, we investigate how to provide timely computational services in VCC systems. In particular, we seek to minimize the deadline violation probability given a set of tasks to be executed in vehicular clouds. Due to the uncertainty of vehicle movements, the task replication methodology is leveraged which allows one task to be executed by several vehicles, and thus trading computational resources for delay reduction. The optimal task replication policy is of key interest. We first formulate the problem as a finite-horizon sampled-time Markov decision problem and obtain the optimal policy by value iterations. To conquer the complexity issue, we propose the balanced-task-assignment (BETA) policy which is proved optimal and has a clear structure: it always assigns the task with the minimum number of replicas. Moreover, a tight closed-form performance upper bound for the BETA policy is derived, which indicates that the deadline violation probability follows the Rayleigh distribution approximately. Applying the vehicle speed-density relationship in the traffic flow theory, we find that vehicle mobility benefits VCC systems more compared with road traffic systems, by showing that the optimum vehicle speed to minimize the deadline violation probability is larger than the critical vehicle speed in traffic theory which maximizes traffic flow efficiency.

A Dynamic Replica Factor Calculator for Weighted Dynamic Replication Management in Cloud Storage Systems

Data replication is a widely-used technique in cloud storage systems to ensure availability of data. In static replication policies, the number of replicas to be created is decided statically at the time of cloud system setup. The replication of data needs to be dynamic considering changing pattern in user request and storage capacity. This will ensure that storage space is not wasted by keeping unnecessary replicas for least used data. In this work, a strategy is proposed to dynamically set the replica factor for each data item considering the popularity of data, its current replication factor and the number of active nodes present in the cloud storage. The proposed strategy is implemented in Hadoop Distributed File System (HDFS) and the experimental results shows that proposed method maintains an optimal number of replicas for each data item based on its popularity, without compromising on the availability constraint of cloud storage.

Do economics journals foster replicable research?

Watch the VIDEO here.Replications are pivotal for the credibility of empirical economics. It is widely recognized in economics that replication studies are a necessary condition for scientific integrity. Alarmingly, several studies indicate that a significant share of empirical economics research cannot be replicated. At the same time, the awareness among researchers, that empirically-based research is often based on shaky grounds, has increased in the last years. It becomes more and more evident that there is a need for more replications in economics to regain trust and credibility in empirical economics research.Though established scholarly journals have adopted replication policies in recent years, replication activities only slightly increased. Against this background our talk investigates if and how journals in economics foster replicable research. For this purpose, we will address two aspects:Journals’ data policies and their effective enforcement in economics: The first part of our talk presents the findings of a new study, in which we evaluated almost 600 articles published in 37 well-regarded journals with a data availability policy. First, we highlight the share of articles that fall under the data policy. Subsequently, the talk contrasts for how many of these data-based articles replication files were available in journals’ data archives and/or the supplemental information section of the article. Moreover, the exact requirements of journals’ data policies have been contrasted to the replication files available on journals’ web pages (respectively in their data archives).We developed a ‘compliance rate’ for each journal in our study. The higher the compliance rate the more do journals enforce their data policy. In the first part of our talk, we also discuss the question whether voluntary data policies are effective in fostering replicable research. For this purpose, we compare the compliance rate of journals with a voluntary data policy to their mandatory counterparts.Journals as publication outlets for replication studies: Though researchers agree that replication studies are needed to regain trust and credibility in empirical economic research, replication activities only slightly increased in recent years. One reason for that finding can be that in the current system replicating other people’s result does not progresses researcher’s career. Another reason is the paucity of publication outlets for such replication studies. In this part of our talk we also discuss whether established journals should implement replication sections or whether a journal which is entirely dedicated to replication, would be a better way to foster the publication of replication studies. As a showcase we will briefly introduce the newly founded “International Journal for Re-Views in Empirical Economics” (IREE).To conclude, we sketch the current and potential future developments in economics when it comes to reproducible research.

Fault-Tolerant Dynamic Task Mapping and Scheduling for Network-on-Chip-Based Multicore Platform

In Network-on-Chip (NoC)-based multicore systems, task allocation and scheduling are known to be important problems, as they affect the performance of applications in terms of energy consumption and timing. Advancement of deep submicron technology has made it possible to scale the transistor feature size to the nanometer range, which has enabled multiple processing elements to be integrated onto a single chip. On the flipside, it has made the integrated entities on the chip more susceptible to different faults. Although a significant amount of work has been done in the domain of fault-tolerant mapping and scheduling, existing algorithms either precompute reconfigured mapping solutions at design time while anticipating fault(s) scenarios or adopt a hybrid approach wherein a part of the fault mitigation strategy relies on the design-time solution. The complexity of the problem rises further for real-time dynamic systems where new applications can arrive in the multicore platform at any time instant. For real-time systems, the validity of computation depends both on the correctness of results and on temporal constraint satisfaction. This article presents an improved fault-tolerant dynamic solution to the integrated problem of application mapping and scheduling for NoC-based multicore platforms. The developed algorithm provides a unified mapping and scheduling method for real-time systems focusing on meeting application deadlines and minimizing communication energy. A predictive model has been used to determine the failure-prone cores in the system for which a fault-tolerant resource allocation with task redundancy has been performed. By selectively using a task replication policy, the reliability of the application, executing on a given NoC platform, is improved. A detailed evaluation of the performance of the proposed algorithm has been conducted for both real and synthetic applications. When compared with other fault-tolerant algorithms reported in the literature, performance of the proposed algorithm shows an average reduction of 56.95% in task re-execution time overhead and an average improvement of 31% in communication energy. Further, for time-constrained tasks, deadline satisfaction has also been achieved for most of the test cases by the developed algorithm, whereas the techniques reported in the literature failed to meet deadline in about 45% test cases.

CAIS: A Copy Adjustable Incentive Scheme in Community-Based Socially Aware Networking

Socially aware networking (SAN) is a new communication paradigm, in which the social characteristics of mobile nodes are exploited to improve the performance of data distribution. In SAN, mobile carriers may exhibit selfish behaviors and refuse to relay messages for others for various reasons, such as limited resources (e.g., buffer, energy, and bandwidth) or social relationships. Several incentive schemes have recently been investigated to stimulate selfish users for cooperation in data forwarding. However, a majority of the existing methods have not fully studied nodes' social relationships in their selfish behaviors. In this paper, we propose a copy adjustable incentive scheme (CAIS), which adopts the virtual credit concept to stimulate selfish nodes to cooperate in data forwarding. In CAIS, we consider a network in which the nodes are divided into certain communities based on their social relationships. Then, we apply two types of credits, i.e., social credit and nonsocial credit, to reward the nodes when they relay data for other nodes inside their community or outsiders, respectively. Based on our mechanism, the number of messages a node can replicate to other nodes is adjusted according to its cooperation level and earned credits. To further improve the performance of CAIS, a single-copy data replication policy is employed, which manages the credit distribution of each node according to its available resources. The results of our extensive experiments using both synthetic and trace-driven simulations illustrate that CAIS copes well with node selfishness in community-based networks and outperforms other benchmark protocols with high data delivery ratio, low communication overhead, and short data delivery latency.

Leveraged Funds: Robust Replication and Performance Evaluation

Leveraged and inverse ETFs seek a daily return equal to a multiple of an index' return, an objective that requires continuous portfolio rebalancing. The resulting trading costs create a tradeoff between tracking error, which controls the short-term correlation with the index, and excess return (or tracking difference) - the long-term deviation from the leveraged index' performance. With proportional trading costs, the optimal replication policy is robust to the index' dynamics. A summary of a fund's performance is the implied spread, equal to the product of tracking error and excess return, rescaled for leverage and average volatility. The implied spread is insensitive to the benchmark's risk premium and offers a tool to compare the performance of funds tracking the same index with different factors and tracking errors.

How Are We Doing? Data Access and Replication in Political Science

ABSTRACTData access and research transparency (DA-RT) is a growing concern for the discipline. Technological advances have greatly reduced the cost of sharing data, enabling full replication archives consisting of data and code to be shared on individual websites, as well as journal archives and institutional data repositories. But how do we ensure that scholars take advantage of these resources to share their replication archives? Moreover, are the costs of research transparency borne by individuals or by journals? This article assesses the impact of journal replication policies on data availability and finds that articles published in journals with mandatory provision policies are 24 times more likely to have replication materials available than articles those with no requirements.

Effect of selective replication strategy and dynamic admission control to QoS management in video on demand systems

Abstract Several solutions have been proposed till date to achieve effective replication policies and admission control methods that can ensure a high Quality of Service (QoS) to Video on Demand (VoD) systems. In this paper, we have chosen to study admission control and a replication strategy in a video on demand system called Feedback Control Architecture for Distributed Multimedia Systems (FCA-DMS). The proposed approaches are founded on (i) the supervision and auto-adaptation of the system load, (ii) continued knowledge of the network and of the video servers’ workload, and (iii) the development of a dynamic replication strategy of segments for popular video, that remade to replica-placement problem by discussing the problem of the replicas number to create (Replication Degree). The workload of video servers and of network are calculated by a QoS controller. The proposed approach takes into account of replica detail and offers a replica removal policy to avoid waste and for a good management of storage space. The goal is to maintain an updated broad view of the system’s behaviour from the collected measurements to databases of available videos and databases of available video servers. To demonstrate the efficiency and feasibility of our approaches and validate the results obtained, we conducted a series of experiments based on a simulator that we implemented.

Replication Policy of Real-Time Distributed System for Cloud Computing

Recently, cloud computing has been widely used for the purpose of protecting client data on the Internet [A. Weiss, Computing in the clouds, netWorker11 (2007) 16–25; M. Armbrust et al., Above the clouds: A Berkeley view of cloud computing, Technical Report UCV/EECS-2009-28, University of California at Berkeley (2009)]. But when a client receives network service, response time may be slow because the data center is located in a remote place. In order to solve the problem, real-time distributed systems for cloud computing has been proposed [M. Okuno, D. Ito, H. Miyamoto, H. Aoki, Y. Tsushima and T. Yazaki, A study on distributed information and communication processing architecture for next generation cloud system, IEICE Tech. Report109(A48) (2010) 241–246; M. Okuno, S. Tsutsumi and T. Yazaki, A study of high available distributed network processing technique for next generation cloud system, IEICE Tech. Report111(8) (2011) 25–30; S. Yamada, J. Marukawa, D. Ishii, S. Okamoto and N. Yamanaka, A study of parallel transmission technique with GMPLS in intelligent cloud network, IEICE Tech. Report109(455) (2010) 51–56]. The cloud computing system consists of some intelligent nodes as well as a data center. The data center manages all client data. The intelligent node provides client service near clients. It enables to provide client service at short response time [M. Okuno, D. Ito, H. Miyamoto, H. Aoki, Y. Tsushima and T. Yazaki, A study on distributed information and communication processing architecture for next generation cloud system, IEICE Tech. Report109(448) (2010) 241–246]. We considered the reliability model of distributed information processing for cloud computing, derived cost effectiveness and discussed the optimal replication interval to minimize it [M. Kimura, M. Imaizumi and T. Nakagawa, Reliability modeling of distributed information processing for cloud computing, in Proc. 20th ISSAT Int. Conf. Reliability and Quality in Design (2014), pp. 183–187]. Authors had dealt with the server system with one failure mode. In this paper, we consider the reliability model of a real-time distributed system with n intelligent nodes and formulate a stochastic model of the server system with n intelligent nodes for changing the other normal intelligent node at failure. We derive the expected numbers of the replication and of updating the client data. Further, we derive the expected cost and discuss an optimal replication interval to minimize it. Next, we derive the cost effectiveness and discuss an optimal number of intelligent nodes to minimize it.

When should I make preservation copies of myself?

We investigate how different replication policies ranging from least aggressive to most aggressive affect the level of preservation achieved by autonomic processes used by web objects (WOs). Based on simulations of small-world graphs of WOs created by the Unsupervised Small-World algorithm, we report quantitative and qualitative results for graphs ranging in order from 10 to 5000 WOs. Our results show that a moderately aggressive replication policy makes the best use of distributed host resources by not causing spikes in CPU resources nor spikes in network activity while meeting preservation goals. We examine different approaches that WOs can communicate with each other and determine the how long it would take for a message from one WO to reach a specific WO, or all WOs.

Impact of MapReduce Policies on Job Completion Reliability and Job Energy Consumption

Recently, MapReduce has been widely employed by many companies/organizations to tackle data-intensive problems over a large-scale MapReduce cluster. To solve machine/node failure which is inevitable in a MapReduce cluster, MapReduce employs several policies, such as input-data replication and intermediate-data replication policies. To speed up job execution, MapReduce allows reduce tasks to early fetch their required intermediate data. However, the impact of these policy combinations on the job completion reliability (JCR for short) and job energy consumption (JEC for short) of a MapReduce cluster was not clear, where JCR is the reliability with which a MapReduce job can be completed by the cluster, whereas JEC is the energy consumed by the cluster to complete the job. Therefore, in this study, we analyze the JCR and JEC of a MapReduce cluster on four policy combinations (POCs for short) derived from two typical intermediate-data replication policies and two typical reduce-task assignment policies. The four POCs are further compared in extensive scenarios, which not only consider jobs at different scales with various parameters, but also give a MapReduce cluster two extreme parallel execution capabilities and diverse bandwidths. The analytical results enable MapReduce managers to comprehend how these POCs impact the JCR and JEC of a cluster and then select an appropriate POC based on the characteristics of their own MapReduce jobs and clusters.