Cost-intelligent application-specific data layout optimization for parallel file systems

Parallel file systems have been developed in recent years to ease the I/O bottleneck of high-end computing system. These advanced file systems offer several data layout strategies in order to meet the performance goals of specific I/O workloads. However, while a layout policy may perform well on some I/O workload, it may not perform as well for another. Peak I/O performance is rarely achieved due to the complex data access patterns. Data access is application dependent. In this study, a cost-intelligent data access strategy based on the application-specific optimization principle is proposed. This strategy improves the I/O performance of parallel file systems. We first present examples to illustrate the difference of performance under different data layouts. By developing a cost model which estimates the completion time of data accesses in various data layouts, the layout can better match the application. Static layout optimization can be used for applications with dominant data access patterns, and dynamic layout selection with hybrid replications can be used for applications with complex I/O patterns. Theoretical analysis and experimental testing have been conducted to verify the proposed cost-intelligent layout approach. Analytical and experimental results show that the proposed cost model is effective and the application-specific data layout approach can provide up to a 74% performance improvement for data-intensive applications.     

The MOSIX Direct File System Access Method for Supporting Scalable Cluster File Systems

MOSIX is a cluster management system that supports preemptive process migration. This paper presents the MOSIX Direct File System Access (DFSA), a provision that can improve the performance of cluster file systems by allowing a migrated process to directly access files in its current location. This capability, when combined with an appropriate file system, could substantially increase the I/O performance and reduce the network congestion by migrating an I/O intensive process to a file server rather than the traditional way of bringing the file's data to the process. DFSA is suitable for clusters that manage a pool of shared disks among multiple machines. With DFSA, it is possible to migrate parallel processes from a client node to file servers for parallel access to different files. Any consistent file system can be adjusted to work with DFSA. To test its performance, we developed the MOSIX File-System (MFS) which allows consistent parallel operations on different files. The paper describes DFSA and presents the performance of MFS with and without DFSA.     

ORTHRUS: a lightweighted block-level cloud storage system

Taking advantage of distributed storage technology and virtualization technology, cloud storage systems provide virtual machine clients customizable storage service. They can be divided into two types: distributed file system and block level storage system. There are two disadvantages in existing block level storage system: Firstly, Some of them are tightly coupled with their cloud computing environments. As a result, it’s hard to extend them to support other cloud computing platforms; Secondly, The bottleneck of volume server seriously affects the performance and reliability of the whole system. In this paper we present a lightweighted block-level storage system for clouds—ORTHRUS, based on virtualization technology. We first design the architecture with multiple volume servers and its workflows, which can improve system performance and avoid the problem. Secondly, we propose a Listen-Detect-Switch mechanism for ORTHRUS to deal with contingent volume servers’ failure. At last we design a strategy that dynamically balances load between multiple volume servers. We characterize machine capability and load quantity with black box model, and implement the dynamic load balance strategy which is based on genetic algorithm. Extensive experimental results show that the aggregated I/O throughputs of ORTHRUS are significantly improved (approximately two times of that in Orthrus), and both I/O throughputs and IOPS are also remarkably improved (about 1.8 and 1.2 times, respectively) by our dynamic load balance strategy.     

A resource query interface for network-aware applications

Networked systems provide a cost-effective platform for parallel computing, but the applications have to deal with the changing availability of computation and communication resources. Network-awareness is a recent attempt to bridge the gap between the realities of networks and the demands of applications. Network-aware applications obtain information about their execution environment and dynamically adapt to enhance their performance. Adaptation is especially important for synchronous parallel applications because a single busy communication link can become the bottleneck and degrade overall performance dramatically. This paper presents Remos, a uniform API that allows applications to obtain relevant network information, and reports on the development of parallel applications in this environment. The challenges in defining a uniform interface include network heterogeneity, diversity and variability in network traffic, and resource sharing in the network and even inside an application. The first implementation of the Remos interface uses SNMP to monitor IP-based networks. This paper reports on our methodology for developing adaptive parallel applications for high-speed networks with Remos and presents experimental results using applications generated by the Fx parallelizing compiler. The results highlight the importance and effectiveness of adaptive parallel computing. This revised version was published online in July 2006 with corrections to the Cover Date.     

Multi-Scale Analysis of the European Airspace Using Network Community Detection

We show that the European airspace can be represented as a multi-scale traffic network whose nodes are airports, sectors, or navigation points and links are defined and weighted according to the traffic of flights between the nodes. By using a unique database of the air traffic in the European airspace, we investigate the architecture of these networks with a special emphasis on their community structure. We propose that unsupervised network community detection algorithms can be used to monitor the current use of the airspace and improve it by guiding the design of new ones. Specifically, we compare the performance of several community detection algorithms, both with fixed and variable resolution, and also by using a null model which takes into account the spatial distance between nodes, and we discuss their ability to find communities that could be used to define new control units of the airspace.     

DataStager: scalable data staging services for petascale applications

Known challenges for petascale machines are that (1) the costs of I/O for high performance applications can be substantial, especially for output tasks like checkpointing, and (2) noise from I/O actions can inject undesirable delays into the runtimes of such codes on individual compute nodes. This paper introduces the flexible ‘DataStager’ framework for data staging and alternative services within that jointly address (1) and (2). Data staging services moving output data from compute nodes to staging or I/O nodes prior to storage are used to reduce I/O overheads on applications’ total processing times, and explicit management of data staging offers reduced perturbation when extracting output data from a petascale machine’s compute partition. Experimental evaluations of DataStager on the Cray XT machine at Oak Ridge National Laboratory establish both the necessity of intelligent data staging and the high performance of our approach, using the GTC fusion modeling code and benchmarks running on 1000+ processors.     

Adaptive hybrid storage systems leveraging SSDs and HDDs in HPC cloud environments

Cloud computing should inherently support various types of data-intensive workloads with different storage access patterns. This makes a high-performance storage system in the Cloud an important component. Emerging flash device technologies such as solid state drives (SSDs) are a viable choice for building high performance computing (HPC) cloud storage systems to address more fine-grained data access patterns. However, the bit-per-dollar SSD price is still higher than the prices of HDDs. This study proposes an optimized progressive file layout (PFL) method to leverage the advantages of SSDs in a parallel file system such as Lustre so that small file I/O performance can be significantly improved. A PFL can dynamically adjust chunk sizes and stripe patterns according to various I/O traffics. Extensive experimental results show that this approach (i.e. building a hybrid storage system based on a combination of SSDs and HDDs) can actually achieve balanced throughput over mixed I/O workloads consisting of large and small file access patterns.     

DENS: data center energy-efficient network-aware scheduling

In modern data centers, energy consumption accounts for a considerably large slice of operational expenses. The existing work in data center energy optimization is focusing only on job distribution between computing servers based on workload or thermal profiles. This paper underlines the role of communication fabric in data center energy consumption and presents a scheduling approach that combines energy efficiency and network awareness, named DENS. The DENS methodology balances the energy consumption of a data center, individual job performance, and traffic demands. The proposed approach optimizes the tradeoff between job consolidation (to minimize the amount of computing servers) and distribution of traffic patterns (to avoid hotspots in the data center network).     

Supporting high‐performance I/O in QoS‐enabled ORB middleware

To be an effective platform for high‐performance distributed applications, off-the-shelf Object Request Broker (ORB) middleware, such as CORBA, must preserve communication-layer quality of service (QoS) properties both vertically (i.e., network interface ↔ application layer) and horizontally (i.e., end-to-end). However, conventional network interfaces, I/O subsystems, and middleware interoperability protocols are not well-suited for applications that possess stringent throughput, latency, and jitter requirements. It is essential, therefore, to develop vertically and horizontally integrated ORB endsystems that can be (1) configured flexibly to support high-performance network interfaces and I/O subsystems and (2) used transparently by performance-sensitive applications. This paper provides three contributions to research on high-performance I/O support for QoS-enabled ORB middleware. First, we outline the key research challenges faced by high-performance ORB endsystem developers. Second, we describe how our real-time I/O (RIO) subsystem and pluggable protocol framework enables ORB endsystems to preserve high-performance network interface QoS up to applications running on off-the-shelf hardware and software. Third, we illustrate empirically how highly optimized ORB middleware can be integrated with real-time I/O subsystem to reduce latency bounds on communication between high-priority clients without unduly penalizing low-priority and best-effort clients. Our results demonstrate how it is possible to develop ORB endsystems that are both highly flexible and highly efficient. This revised version was published online in July 2006 with corrections to the Cover Date.     

Network slicing to improve multicasting in HPC clusters

In high performance computing (HPC) resources’ extensive experiments are frequently executed. HPC resources (e.g. computing machines and switches) should be able to handle running several experiments in parallel. Typically HPC utilizes parallelization in programs, processing and data. The underlying network is seen as the only non-parallelized HPC component (i.e. no dynamic virtual slicing based on HPC jobs). In this scope we present an approach in this paper to utilize software defined networking (SDN) to parallelize HPC clusters among the different running experiments. We propose to accomplish this through two major components: A passive module (network mapper/remapper) to select for each experiment as soon as it starts the least busy resources in the network, and an SDN-HPC active load balancer to perform more complex and intelligent operations. Active load balancer can logically divide the network based on experiments’ host files. The goal is to reduce traffic to unnecessary hosts or ports. An HPC experiment should multicast, rather than broadcast to only cluster nodes that are used by the experiment. We use virtual tenant network modules in Opendaylight controller to create VLANs based on HPC experiments. In each HPC host, virtual interfaces are created to isolate traffic from the different experiments. The traffic between the different physical hosts that belong to the same experiment can be distinguished based on the VLAN ID assigned to each experiment. We evaluate the new approach using several HPC public benchmarks. Results show a significant enhancement in experiments’ performance especially when HPC cluster experiences running several heavy load experiments simultaneously. Results show also that this multi-casting approach can significantly reduce casting overhead that is caused by using a single cast for all resources in the HPC cluster. In comparison with InfiniBand networks that offer interconnect services with low latency and high bandwidth, HPC services based on SDN can provide two distinguished objectives that may not be possible with InfiniBand: The first objective is the integration of HPC with Ethernet enterprise networks and hence expanding HPC usage to much wider domains. The second objective is the ability to enable users and their applications to customize HPC services with different QoS requirements that fit the different needs of those applications and optimize the usage of HPC clusters.     

Behavior and Performance of Interactive Multi-Player Game Servers

With the recent explosion in deployment of services to large numbers of customers over the Internet and in global services in general, issues related to the architecture of scalable servers are becoming increasingly important. However, our understanding of these types of applications is currently limited, especially on how well they scale to support large numbers of users. One such, novel, commercial class of applications, are interactive, multi-player game servers. Multi-player games are both an important class of commercial applications (in the entertainment industry) and they can be valuable in understanding the architectural requirements of scalable services. They impose requirements on system performance, scalability, and availability, stressing multiple aspects of the system architecture (e.g., compute cycles and network I/O). Recently there has been a lot of interest on client side issues with respect to games. However, there has beenlittle or no work on the server side. In this paper we use a commercial game server to gain insight in this class of applications and the requirements they impose on modern architectures. We find that: (1) In terms of the benchmarking methodology, interactive game servers are very different from scientific workloads. We propose a methodology that deals with the related issues in benchmarking this class of applications. Our methodology bears many similarities with methodologies used in benchmarking online transaction processing (OLTP) systems. (2) Current, sequential game servers can support at most up to a few tens of users (60–100) on existing processors. (3) The bottleneck in the server is both game-related as well as network-related processing (about 50–50). (4) Network bandwidth requirements are not an important issue for the numbers of players we are interested in. (5) The processor achieves a surprisingly low IPC of 0.416.     

A queueing model for predicting message latency in uni-directional <Emphasis Type="Italic">k</Emphasis>-ary <Emphasis Type="Italic">n</Emphasis>-cubes with deterministic routing and non-uniform traffic

The interconnection network is one of the key architectural components in any parallel computer. The distribution of the traffic injected into the network is among the factors that greatly influences network performance. The uniform traffic pattern has been adopted in many existing network performance evaluation studies due to the tractability of the resulting analytical modelling approach. However, many real applications exhibit non-uniform traffic patterns such as hot-spot traffic. K-ary n-cubes have been the mostly widely used in the implementation of practical parallel systems. Extensive research studies have been conducted on the performance modelling and evaluation of these networks. Nonetheless, most of these studies have been confined to uniform traffic distributions and have been based on software simulation. The present paper proposes a new stochastic model to predict message latency in k-ary n-cubes with deterministic routing in the presence of hot-spot traffic. The model has been validated through simulation experiments and has shown a close agreement with simulation results.

Improving the performance of I/O-intensive applications on clusters of workstations

Load balancing in a workstation-based cluster system has been investigated extensively, mainly focusing on the effective usage of global CPU and memory resources. However, if a significant portion of applications running in the system is I/O-intensive, traditional load balancing policies can cause system performance to decrease substantially. In this paper, two I/O-aware load-balancing schemes, referred to as IOCM and WAL-PM, are presented to improve the overall performance of a cluster system with a general and practical workload including I/O activities. The proposed schemes dynamically detect I/O load imbalance of nodes in a cluster, and determine whether to migrate some I/O load from overloaded nodes to other less- or under-loaded nodes. The current running jobs are eligible to be migrated in WAL-PM only if overall performance improves. Besides balancing I/O load, the scheme judiciously takes into account both CPU and memory load sharing in the system, thereby maintaining the same level of performance as existing schemes when I/O load is low or well balanced. Extensive trace-driven simulations for both synthetic and real I/O-intensive applications show that: (1) Compared with existing schemes that only consider CPU and memory, the proposed schemes improve the performance with respect to mean slowdown by up to a factor of 20; (2) When compared to the existing approaches that only consider I/O with non-preemptive job migrations, the proposed schemes achieve improvements in mean slowdown by up to a factor of 10; (3) Under CPU-memory intensive workloads, our scheme improves the performance over the existing approaches that only consider I/O by up to 47.5%. Xiao Qin received the BSc and MSc degrees in computer science from Huazhong University of Science and Technology in 1992 and 1999, respectively. He received the PhD degree in computer science from the University of Nebraska-Lincoln in 2004. Currently, he is an assistant professor in the department of computer science at the New Mexico Institute of Mining and Technology. His research interests include parallel and distributed systems, storage systems, real-time computing, performance evaluation, and fault-tolerance. He served on program committees of international conferences like CLUSTER, ICPP, and IPCCC. During 2000–2001, he was on the editorial board of The IEEE Distributed System Online. He is a member of the IEEE. Hong Jiang received the B.Sc. degree in Computer Engineering in 1982 from Huazhong University of Science and Technology, Wuhan, China; the M.A.Sc. degree in Computer Engineering in 1987 from the University of Toronto, Toronto, Canada; and the PhD degree in Computer Science in 1991 from the Texas A&M University, College Station, Texas, USA. Since August 1991 he has been at the University of Nebraska-Lincoln, Lincoln, Nebraska, USA, where he is Associate Professor and Vice Chair in the Department of Computer Science and Engineering. His present research interests are computer architecture, parallel/distributed computing, computer storage systems and parallel I/O, performance evaluation, middleware, networking, and computational engineering. He has over 70 publications in major journals and international Conferences in these areas and his research has been supported by NSF, DOD and the State of Nebraska. Dr. Jiang is a Member of ACM, the IEEE Computer Society, and the ACM SIGARCH and ACM SIGCOMM. Yifeng Zhu received the B.E. degree in Electrical Engineering from Huazhong University of Science and Technology in 1998 and the M.S. degree in computer science from University of Nebraska Lincoln (UNL) in 2002. Currently he is working towards his Ph.D. degree in the department of computer science and engineering at UNL. His main fields of research interests are parallel I/O, networked storage, parallel scheduling, and cluster computing. He is a student member of IEEE. David Swanson received a Ph.D. in physical (computational) chemistry at the University of Nebraska-Lincoln (UNL) in 1995, after which he worked as an NSF-NATO postdoctoral fellow at the Technical University of Wroclaw, Poland, in 1996, and subsequently as a National Research Council Research Associate at the Naval Research Laboratory in Washington, DC, from 1997–1998. In early 1999 he returned to UNL where he has coordinated the Research Computing Facility and currently serves as an Assistant Research Professor in the Department of Computer Science and Engineering. The Office of Naval Research, the National Science Foundation, and the State of Nebraska have supported his research in areas such as large-scale parallel simulation and distributed systems.     

neCODEC: nearline data compression for scientific applications

Advances on multicore technologies lead to processors with tens and soon hundreds of cores in a single socket, resulting in an ever growing gap between computing power and available memory and I/O bandwidths for data handling. It would be beneficial if some of the computing power can be transformed into gains of I/O efficiency, thereby reducing this speed disparity between computing and I/O. In this paper, we design and implement a NEarline data COmpression and DECompression (neCODEC) scheme for data-intensive parallel applications. Several salient techniques are introduced in neCODEC, including asynchronous compression threads, elastic file representation, distributed metadata handling, and balanced subfile distribution. Our performance evaluation indicates that neCODEC can improve the performance of a variety of data-intensive microbenchmarks and scientific applications. Particularly, neCODEC is capable of increasing the effective bandwidth of S3D, a combustion simulation code, by more than 5 times.     

A Framework for Adaptive Cluster Computing Using JavaSpaces

Heterogeneous networked clusters are being increasingly used as platforms for resource-intensive parallel and distributed applications. The fundamental underlying idea is to provide large amounts of processing capacity over extended periods of time by harnessing the idle and available resources on the network in an opportunistic manner. In this paper we present the design, implementation and evaluation of a framework that uses JavaSpaces to support this type of opportunistic adaptive parallel/distributed computing over networked clusters in a non-intrusive manner. The framework targets applications exhibiting coarse grained parallelism and has three key features: (1) portability across heterogeneous platforms, (2) minimal configuration overheads for participating nodes, and (3) automated system state monitoring (using SNMP) to ensure non-intrusive behavior. Experimental results presented in this paper demonstrate that for applications that can be broken into coarse-grained, relatively independent tasks, the opportunistic adaptive parallel computing framework can provide performance gains. Furthermore, the results indicate that monitoring and reacting to the current system state minimizes the intrusiveness of the framework.     

Complex Network Theory Applied to the Growth of Kuala Lumpur’s Public Urban Rail Transit Network

Recently, the number of studies involving complex network applications in transportation has increased steadily as scholars from various fields analyze traffic networks. Nonetheless, research on rail network growth is relatively rare. This research examines the evolution of the Public Urban Rail Transit Networks of Kuala Lumpur (PURTNoKL) based on complex network theory and covers both the topological structure of the rail system and future trends in network growth. In addition, network performance when facing different attack strategies is also assessed. Three topological network characteristics are considered: connections, clustering and centrality. In PURTNoKL, we found that the total number of nodes and edges exhibit a linear relationship and that the average degree stays within the interval [2.0488, 2.6774] with heavy-tailed distributions. The evolutionary process shows that the cumulative probability distribution (CPD) of degree and the average shortest path length show good fit with exponential distribution and normal distribution, respectively. Moreover, PURTNoKL exhibits clear cluster characteristics; most of the nodes have a 2-core value, and the CPDs of the centrality’s closeness and betweenness follow a normal distribution function and an exponential distribution, respectively. Finally, we discuss four different types of network growth styles and the line extension process, which reveal that the rail network’s growth is likely based on the nodes with the biggest lengths of the shortest path and that network protection should emphasize those nodes with the largest degrees and the highest betweenness values. This research may enhance the networkability of the rail system and better shape the future growth of public rail networks.     

Discretionary Caching for I/O on Clusters

I/O bottlenecks are already a problem in many large-scale applications that manipulate huge datasets. This problem is expected to get worse as applications get larger, and the I/O subsystem performance lags behind processor and memory speed improvements. At the same time, off-the-shelf clusters of workstations are becoming a popular platform for demanding applications due to their cost-effectiveness and widespread deployment. Caching I/O blocks is one effective way of alleviating disk latencies, and there can be multiple levels of caching on a cluster of workstations. Previous studies have shown the benefits of caching—whether it be local to a particular node, or a shared global cache across the cluster—for certain applications. However, we show that while caching is useful in some situations, it can hurt performance if we are not careful about what to cache and when to bypass the cache. This paper presents compilation techniques and runtime support to address this problem. These techniques are implemented and evaluated on an experimental Linux/Pentium cluster running a parallel file system. Our results using a diverse set of applications (scientific and commercial) demonstrate the benefits of a discretionary approach to caching for I/O subsystems on clusters, providing as much as 48% savings in overall execution time over indiscriminately caching everything in some applications. Parts of this paper have appeared in the Proceedings of the 3rd IEEE/ACM Symposium on Cluster Computing and the Grid (CCGrid'03). This paper is an extension of these prior results, and includes a more extensive performance evaluation. Murali Vilayannur is a Ph.D. student in the Department of Computer Science and Engineering at The Pennsylvania State University. His research interests are in High-Performance Parallel I/O, File Systems, Virtual Memory Algorithms and Operating Systems. Anand Sivasubramaniam received his B.Tech. in Computer Science from the Indian Institute of Technology, Madras, in 1989, and the M.S. and Ph.D. degrees in Computer Science from the Georgia Institute of Technology in 1991 and 1995 respectively. He has been on the faculty at The Pennsylvania State University since Fall 1995 where he is currently an Associate Professor. Anand's research interests are in computer architecture, operating systems, performance evaluation, and applications for both high performance computer systems and embedded systems. Anand's research has been funded by NSF through several grants, including the CAREER award, and from industries including IBM, Microsoft and Unisys Corp. He has several publications in leading journals and conferences, and is on the editorial board of IEEE Transactions on Computers and IEEE Transactions on Parallel and Distributed Systems. He is a recipient of the 2002 IBM Faculty Award. Anand is a member of the IEEE, IEEE Computer Society, and ACM. Mahmut Kandemir received the B.Sc. and M.Sc. degrees in control and computer engineering from Istanbul Technical University, Istanbul, Turkey, in 1988 and 1992, respectively. He received the Ph.D. from Syracuse University, Syracuse, New York in electrical engineering and computer science, in 1999. He has been an assistant professor in the Computer Science and Engineering Department at the Pennsylvania State University since August 1999. His main research interests are optimizing compilers, I/O intensive applications, and power-aware computing. He is a member of the IEEE and the ACM. Rajeev Thakur is a Computer Scientist in the Mathematics and Computer Science Division at Argonne National Laboratory. He received a B.E. from the University of Bombay, India, in 1990, M.S. from Syracuse University in 1992, and Ph.D. from Syracuse University in 1995, all in computer engineering. His research interests are in the area of high-performance computing in general and high-performance networking and I/O in particular. He was a member of the MPI Forum and participated actively in the definition of the I/O part of the MPI-2 standard. He is the author of a widely used, portable implementation of MPI-IO, called ROMIO. He is also a co-author of the book “Using MPI-2: Advanced Features of the Message Passing Interface” published by MIT Press. Robert Ross received his Ph.D. in Computer Engineering from Clemson University in 2000. He is now an Assistant Scientist in the Mathematics and Computer Science Division at Argonne National Laboratory. His research interests are in message passing and storage systems for high performance computing environments. He is the primary author and lead developer for the Parallel Virtual File System (PVFS), a parallel file system for Linux clusters. Current projects include the ROMIO MPI-IO implementation, PVFS, PVFS2, and the MPICH2 implementation of the MPI message passing interface.     

LSMAC vs. LSNAT: Scalable cluster‐based Web servers

Server scalability is more important than ever in today's client/server dominated network environments. Recently, researchers have begun to consider cluster-based computers using commodity hardware as an alternative to expensive specialized hardware for building scalable Web servers. In this paper, we present performance results comparing two cluster-based Web servers based on different server architectures: OSI layer two dispatching (LSMAC) and OSI layer three dispatching (LSNAT). Both cluster-based server systems were implemented as application-space programs running on commodity hardware in contrast to other, similar, solutions which require specialized hardware/software. We point out the advantages and disadvantages of both systems. We also identify when servers should be clustered and when clustering will not improve performance. This revised version was published online in July 2006 with corrections to the Cover Date.     

A metric of influential spreading during contagion dynamics through the air transportation network

The spread of infectious diseases at the global scale is mediated by long-range human travel. Our ability to predict the impact of an outbreak on human health requires understanding the spatiotemporal signature of early-time spreading from a specific location. Here, we show that network topology, geography, traffic structure and individual mobility patterns are all essential for accurate predictions of disease spreading. Specifically, we study contagion dynamics through the air transportation network by means of a stochastic agent-tracking model that accounts for the spatial distribution of airports, detailed air traffic and the correlated nature of mobility patterns and waiting-time distributions of individual agents. From the simulation results and the empirical air-travel data, we formulate a metric of influential spreading--the geographic spreading centrality--which accounts for spatial organization and the hierarchical structure of the network traffic, and provides an accurate measure of the early-time spreading power of individual nodes.     

A Coarray Fortran implementation to support data-intensive application development

In this paper, we describe our experiences in implementing and applying Coarray Fortran (CAF) for the development of data-intensive applications in the domain of Oil and Gas exploration. The successful porting of reverse time migration (RTM), a data-intensive algorithm and one of the largest uses of computational resources in seismic exploration, is described, and results are presented demonstrating that the CAF implementation provides comparable performance to an equivalent MPI version. We then discuss further language extensions for supporting scalable parallel I/O operating on the massive data sets that are typical of applications used in seismic exploration.