Alexandru Iosup

Performance Analysis of Cloud Computing Services for Many-Tasks Scientific Computing

Cloud computing is an emerging commercial infrastructure paradigm that promises to eliminate the need for maintaining expensive computing facilities by companies and institutes alike. Through the use of virtualization and resource time sharing, clouds serve with a single set of physical resources a large user base with different needs. Thus, clouds have the potential to provide to their owners the benefits of an economy of scale and, at the same time, become an alternative for scientists to clusters, grids, and parallel production environments. However, the current commercial clouds have been built to support web and small database workloads, which are very different from typical scientific computing workloads. Moreover, the use of virtualization and resource time sharing may introduce significant performance penalties for the demanding scientific computing workloads. In this work, we analyze the performance of cloud computing services for scientific computing workloads. We quantify the presence in real scientific computing workloads of Many-Task Computing (MTC) users, that is, of users who employ loosely coupled applications comprising many tasks to achieve their scientific goals. Then, we perform an empirical evaluation of the performance of four commercial cloud computing services including Amazon EC2, which is currently the largest commercial cloud. Last, we compare through trace-based simulation the performance characteristics and cost models of clouds and other scientific computing platforms, for general and MTC-based scientific computing workloads. Our results indicate that the current clouds need an order of magnitude in performance improvement to be useful to the scientific community, and show which improvements should be considered first to address this discrepancy between offer and demand.

TRIBLER: a social-based peer-to-peer system

Most current peer‐to‐peer (P2P) file‐sharing systems treat their users as anonymous, unrelated entities, and completely disregard any social relationships between them. However, social phenomena such as friendship and the existence of communities of users with similar tastes or interests may well be exploited in such systems in order to increase their usability and performance. In this paper we present a novel social‐based P2P file‐sharing paradigm that exploits social phenomena by maintaining social networks and using these in content discovery, content recommendation, and downloading. Based on this paradigms main concepts such as taste buddies and friends, we have designed and implemented the TRIBLER P2P file‐sharing system as a set of extensions to BitTorrent. We present and discuss the design of TRIBLER, and we show evidence that TRIBLER enables fast content discovery and recommendation at a low additional overhead, and a significant improvement in download performance. Copyright

A Performance Analysis of EC2 Cloud Computing Services for Scientific Computing

Cloud Computing is emerging today as a commercial infrastructure that eliminates the need for maintaining expensive computing hardware. Through the use of virtualization, clouds promise to address with the same shared set of physical resources a large user base with different needs. Thus, clouds promise to be for scientists an alternative to clusters, grids, and supercomputers. However, virtualization may induce significant performance penalties for the demanding scientific computing workloads. In this work we present an evaluation of the usefulness of the current cloud computing services for scientific computing. We analyze the performance of the Amazon EC2 platform using micro-benchmarks and kernels. While clouds are still changing, our results indicate that the current cloud services need an order of magnitude in performance improvement to be useful to the scientific community.

On the Performance Variability of Production Cloud Services

Cloud computing is an emerging infrastructure paradigm that promises to eliminate the need for companies to maintain expensive computing hardware. Through the use of virtualization and resource time-sharing, clouds address with a single set of physical resources a large user base with diverse needs. Thus, clouds have the potential to provide their owners the benefits of an economy of scale and, at the same time, become an alternative for both the industry and the scientific community to self-owned clusters, grids, and parallel production environments. For this potential to become reality, the first generation of commercial clouds need to be proven to be dependable. In this work we analyze the dependability of cloud services. Towards this end, we analyze long-term performance traces from Amazon Web Services and Google App Engine, currently two of the largest commercial clouds in production. We find that the performance of about half of the cloud services we investigate exhibits yearly and daily patterns, but also that most services have periods of especially stable performance. Last, through trace-based simulation we assess the impact of the variability observed for the studied cloud services on three large-scale applications, job execution in scientific computing, virtual goods trading in social networks, and state management in social gaming. We show that the impact of performance variability depends on the application, and give evidence that performance variability can be an important factor in cloud provider selection.

Procedural content generation for games: A survey

Hundreds of millions of people play computer games every day. For them, game content—from 3D objects to abstract puzzles—plays a major entertainment role. Manual labor has so far ensured that the quality and quantity of game content matched the demands of the playing community, but is facing new scalability challenges due to the exponential growth over the last decade of both the gamer population and the production costs. Procedural Content Generation for Games (PCG-G) may address these challenges by automating, or aiding in, game content generation. PCG-G is difficult, since the generator has to create the content, satisfy constraints imposed by the artist, and return interesting instances for gamers. Despite a large body of research focusing on PCG-G, particularly over the past decade, ours is the first comprehensive survey of the field of PCG-G. We first introduce a comprehensive, six-layered taxonomy of game content: bits, space, systems, scenarios, design, and derived. Second, we survey the methods used across the whole field of PCG-G from a large research body. Third, we map PCG-G methods to game content layers; it turns out that many of the methods used to generate game content from one layer can be used to generate content from another. We also survey the use of methods in practice, that is, in commercial or prototype games. Fourth and last, we discuss several directions for future research in PCG-G, which we believe deserve close attention in the near future.

The Failure Trace Archive: Enabling Comparative Analysis of Failures in Diverse Distributed Systems

With the increasing functionality and complexity of distributed systems, resource failures are inevitable. While numerous models and algorithms for dealing with failures exist, the lack of public trace data sets and tools has prevented meaningful comparisons. To facilitate the design, validation, and comparison of fault-tolerant models and algorithms, we have created the Failure Trace Archive (FTA) as an online public repository of availability traces taken from diverse parallel and distributed systems. Our main contributions in this study are the following. First, we describe the design of the archive, in particular the rationale of the standard FTA format, and the design of a toolbox that facilitates automated analysis of trace data sets. Second, applying the toolbox, we present a uniform comparative analysis with statistics and models of failures in nine distributed systems. Third, we show how different interpretations of these data sets can result in different conclusions. This emphasizes the critical need for the public availability of trace data and methods for their analysis.

How are Real Grids Used? The Analysis of Four Grid Traces and Its Implications

The grid computing vision promises to provide the needed platform for a new and more demanding range of applications. For this promise to become true, a number of hurdles, including the design and deployment of adequate resource management and information services, need to be overcome. In this context, understanding the characteristics of real grid workloads is a crucial step for improving the quality of existing grid services, and in guiding the design of new solutions. Towards this goal, in this work we present the characteristics of traces of four real grid environments, namely LCG, Grid3, and TeraGrid, which are among the largest production grids currently deployed, and the DAS, which is a research grid. We focus our analysis on virtual organizations, on users, and on individual jobs characteristics. We further attempt to quantify the evolution and the performance of the grid systems from which our traces originate. Finally, given the scarcity of the information available for analysis purposes, we discuss the requirements of a new format for grid traces, and we propose the establishment of a virtual center for workload-based grid benchmarking data: the grid workloads archive

C-Meter: A Framework for Performance Analysis of Computing Clouds

Cloud computing has emerged as a new technology that provides large amounts of computing and data storage capacity to its users with a promise of increased scalability, high availability, and reduced administration and maintenance costs. As the use of cloud computing environments increases, it becomes crucial to understand the performance of these environments. So, it is of great importance to assess the performance of computing clouds in terms of various metrics, such as the overhead of acquiring and releasing the virtual computing resources, and other virtualization and network communications overheads. To address these issues, we have designed and implemented C-Meter, which is a portable, extensible, and easy-to-use framework for generating and submitting test workloads to computing clouds. In this paper, first we state the requirements for frameworks to assess the performance of computing clouds. Then, we present the architecture of the C-Meter framework and discuss several cloud resource management alternatives. Finally, we present ourearly experiences with C-Meter in Amazon EC2. We show how C-Meter can be used for assessing the overhead of acquiring and releasing the virtual computing resources, for comparing different configurations, and for evaluating different scheduling algorithms.

The performance of bags-of-tasks in large-scale distributed systems

Ever more scientists are employing large-scale distributed systems such as grids for their computational work, instead of tightly coupled high-performance computing systems. However, while these distributed systems are more cost-effective, their heterogeneity in terms of hardware, software, and systems administration, and the lack of accurate resource information leads to inefficient scheduling. In addition, and in contrast to the workloads of tightly coupled high-performance computing systems, a large part of the workloads submitted to these distributed systems consists of large sets (bags) of sequential tasks. Therefore, a realistic performance analysis of scheduling bags-of-tasks in large-scale distributed systems is important. Towards this end, we introduce in this paper a realistic workload model for bags-of-tasks, and we explore through trace-based simulations the design space of scheduling bags-of-tasks. Finally, we identify three new scheduling policies that use only inaccurate information when scheduling, and we compare them against known classes of proposed scheduling policies.

Dynamic Resource Provisioning in Massively Multiplayer Online Games

Todays Massively Multiplayer Online Games (MMOGs) can include millions of concurrent players spread across the world and interacting with each other within a single session. Faced with high resource demand variability and with misfit resource renting policies, the current industry practice is to overprovision for each game tens of self-owned data centers, making the market entry affordable only for big companies. Focusing on the reduction of entry and operational costs, we investigate a new dynamic resource provisioning method for MMOG operation using external data centers as low-cost resource providers. First, we identify in the various types of player interaction a source of short-term load variability, which complements the long-term load variability due to the size of the player population. Then, we introduce a combined MMOG processor, network, and memory load model that takes into account both the player interaction type and the population size. Our model is best used for estimating the MMOG resource demand dynamically, and thus, for dynamic resource provisioning based on the game world entity distribution. We evaluate several classes of online predictors for MMOG entity distribution and propose and tune a neural network-based predictor to deliver good accuracy consistently under real-time performance constraints. We assess using trace-based simulation the impact of the data center policies on the quality of resource provisioning. We find that the dynamic resource provisioning can be much more efficient than its static alternative even when the external data centers are busy, and that data centers with policies unsuitable for MMOGs are penalized by our dynamic resource provisioning method. Finally, we present experimental results showing the real-time parallelization and load balancing of a real game prototype using data center resources provisioned using our method and show its advantage against a rudimentary client threshold approach.