Fernando Chirigati

noWorkflow: Capturing and Analyzing Provenance of Scripts

We propose noWorkflow, a tool that transparently captures provenance of scripts and enables reproducibility. Unlike existing approaches, noWorkflow is non-intrusive and does not require users to change the way they work --- users need not wrap their experiments in scientific workflow systems, install version control systems, or instrument their scripts. The tool leverages Software Engineering techniques, such as abstract syntax tree analysis, reflection, and profiling, to collect different types of provenance, including detailed information about the underlying libraries. We describe how noWorkflow captures multiple kinds of provenance and the different classes of analyses it supports: graph-based visualization; differencing over provenance trails; and inference queries.

Exploring many task computing in scientific workflows

One of the main advantages of using a scientific workflow management system (SWfMS) to orchestrate data flows among scientific activities is to control and register the whole workflow execution. The execution of activities within a workflow with high performance computing (HPC) presents challenges in SWfMS execution control. Current solutions leave the scheduling to the HPC queue system. Since the workflow execution engine does not run on remote clusters, SWfMS are not aware of the parallel strategy of the workflow execution. Consequently, remote execution control and provenance registry of the parallel activities is very limited from the SWfMS side. This work presents a set of components to be included on the workflow specification of any SWMfS to control parallelization of activities as MTC. In addition, these components can gather provenance data during remote workflow execution. Through these MTC components, the parallelization strategy can be registered and reused, and provenance data can be uniformly queried. We have evaluated our approach by performing parameter sweep parallelization in solving the incompressible 3D Navier-Stokes equations. Experimental results show the performance gains with the additional benefits of distributed provenance support.

Chiron: a parallel engine for algebraic scientific workflows

Large‐scale scientific experiments based on computer simulations are typically modeled as scientific workflows, which eases the chaining of different programs. These scientific workflows are defined, executed, and monitored by scientific workflow management systems (SWfMS). As these experiments manage large amounts of data, it becomes critical to execute them in high‐performance computing environments, such as clusters, grids, and clouds. However, few SWfMS provide parallel support. The ones that do so are usually labor‐intensive for workflow developers and have limited primitives to optimize workflow execution. To address these issues, we developed workflow algebra to specify and enable the optimization of parallel execution of scientific workflows. In this paper, we show how the workflow algebra is efficiently implemented in Chiron, an algebraic based parallel scientific workflow engine. Chiron has a unique native distributed provenance mechanism that enables runtime queries in a relational database. We developed two studies to evaluate the performance of our algebraic approach implemented in Chiron; the first study compares Chiron with different approaches, whereas the second one evaluates the scalability of Chiron. By analyzing the results, we conclude that Chiron is efficient in executing scientific workflows, with the benefits of declarative specification and runtime provenance support. Copyright

The more the merrier: efficient multi-source graph traversal

Graph analytics on social networks, Web data, and communication networks has been widely used in a plethora of applications. Many graph analytics algorithms are based on breadth-first search (BFS) graph traversal, which is not only time-consuming for large datasets but also involves much redundant computation when executed multiple times from different start vertices. In this paper, we propose Multi-Source BFS (MS-BFS), an algorithm that is designed to run multiple concurrent BFSs over the same graph on a single CPU core while scaling up as the number of cores increases. MS-BFS leverages the properties of small-world networks, which apply to many real-world graphs, and enables efficient graph traversal that: (i) shares common computation across concurrent BFSs; (ii) greatly reduces the number of random memory accesses; and (iii) does not incur synchronization costs. We demonstrate how a real graph analytics application---all-vertices closeness centrality---can be efficiently solved with MS-BFS. Furthermore, we present an extensive experimental evaluation with both synthetic and real datasets, including Twitter and Wikipedia, showing that MS-BFS provides almost linear scalability with respect to the number of cores and excellent scalability for increasing graph sizes, outperforming state-of-the-art BFS algorithms by more than one order of magnitude when running a large number of BFSs.

ReproZip: Computational Reproducibility With Ease

We present ReproZip, the recommended packaging tool for the SIGMOD Reproducibility Review. ReproZip was designed to simplify the process of making an existing computational experiment reproducible across platforms, even when the experiment was put together without reproducibility in mind. The tool creates a self-contained package for an experiment by automatically tracking and identifying all its required dependencies. The researcher can share the package with others, who can then use ReproZip to unpack the experiment, reproduce the findings on their favorite operating system, as well as modify the original experiment for reuse in new research, all with little effort. The demo will consist of examples of non-trivial experiments, showing how these can be packed in a Linux machine and reproduced on different machines and operating systems. Demo visitors will also be able to pack and reproduce their own experiments.

Data Polygamy: The Many-Many Relationships among Urban Spatio-Temporal Data Sets

The increasing ability to collect data from urban environments, coupled with a push towards openness by governments, has resulted in the availability of numerous spatio-temporal data sets covering diverse aspects of a city. Discovering relationships between these data sets can produce new insights by enabling domain experts to not only test but also generate hypotheses. However, discovering these relationships is difficult. First, a relationship between two data sets may occur only at certain locations and/or time periods. Second, the sheer number and size of the data sets, coupled with the diverse spatial and temporal scales at which the data is available, presents computational challenges on all fronts, from indexing and querying to analyzing them. Finally, it is non-trivial to differentiate between meaningful and spurious relationships. To address these challenges, we propose Data Polygamy, a scalable topology-based framework that allows users to query for statistically significant relationships between spatio-temporal data sets. We have performed an experimental evaluation using over 300 spatial-temporal urban data sets which shows that our approach is scalable and effective at identifying interesting relationships.

Towards integrating workflow and database provenance

While there has been substantial work on both database and workflow provenance, the two problems have only been examined in isolation. It is widely accepted that the existing models are incompatible. Database provenance is fine-grained and captures changes to tuples in a database. In contrast, workflow provenance is represented at a coarser level and reflects the functional model of workflow systems, which is stateless--each computational step derives a new artifact. In this paper, we propose a new approach to combine database and workflow provenance. We address the mismatch between the different kinds of provenance by using a temporal model which explicitly represents the database states as updates are applied. We discuss how, under this model, reproducibility is obtained for workflows that manipulate databases, and how different queries that straddle the two provenance traces can be evaluated. We also describe a proof-of-concept implementation that integrates a workflow system and a commercial relational database.

Reproducible experiments on dynamic resource allocation in cloud data centers

In Wolke et al. 1 we compare the efficiency of different resource allocation strategies experimentally. We focused on dynamic environments where virtual machines need to be allocated and deallocated to servers over time. In this companion paper, we describe the simulation framework and how to run simulations to replicate experiments or run new experiments within the framework. HighlightsSimulation and experimentation framework that allows an extensive evaluation of VM allocation strategies.Supports initial VM allocation controllers and dynamic controllers with VM migrations and random VM arrivals/departures.More than 200 time series data sets describing workload in enterprise data centers.The software framework allows for the design of new experiments and the replication of those published in Wolke et al. 1.

Using Explicit Control Processes in Distributed Workflows to Gather Provenance

Distributing workflow tasks among high performance environments involves local processing and remote execution on clusters and grids. This dis-tribution often needs interoperation between heterogeneous workflow definition languages and their corresponding execution machines. A centralized Workflow Management System (WfMS) can be locally controlling the execution of a workflow that needs a grid WfMS to execute a sub-workflow that requires high performance. Workflow specification languages often provide different control-flow execution structures. Moving from one environment to another requires mappings between these languages. Due to heterogeneity, control-flow structures, available in one system, may not be supported in another. In these heterogeneous distributed environments, provenance gathering becomes also heterogeneous. This work presents control-flow modules that aim to be independent from WfMS. By inserting these control-flow modules on the workflow specification, the workflow execution control becomes less dependent of heterogeneous workflow execution engines. In addition, they can be used to gather provenance data both from local and remote execution, thus allowing the same provenance registration on both environments independent of the heterogeneous WfMS. The proposed modules extend the ordinary workflow tasks by providing dynamic behavioral execution control. They were implemented in the VisTrails graphical workflow enactment engine, which offers a flexible infrastructure for provenance gathering.

Evaluating parameter sweep workflows in high performance computing

Scientific experiments based on computer simulations can be defined, executed and monitored using Scientific Workflow Management Systems (SWfMS). Several SWfMS are available, each with a different goal and a different engine. Due to the exploratory analysis, scientists need to run parameter sweep (PS) workflows, which are workflows that are invoked repeatedly using different input data. These workflows generate a large amount of tasks that are submitted to High Performance Computing (HPC) environments. Different execution models for a workflow may have significant differences in performance in HPC. However, selecting the best execution model for a given workflow is difficult due to the existence of many characteristics of the workflow that may affect the parallel execution. We developed a study to show performance impacts of using different execution models in running PS workflows in HPC. Our study contributes by presenting a characterization of PS workflow patterns (the basis for many existing scientific workflows) and its behavior under different execution models in HPC. We evaluated four execution models to run workflows in parallel. Our study measures the performance behavior of small, large and complex workflows among the evaluated execution models. The results can be used as a guideline to select the best model for a given scientific workflow execution in HPC. Our evaluation may also serve as a basis for workflow designers to analyze the expected behavior of an HPC workflow engine based on the characteristics of PS workflows.