Search

Large-scale high performance computing (HPC) systems typically consist of many thousands of CPUs and storage units used by hundreds to thousands of users simultaneously. Applications from large numbers of users have diverse characteristics, such as varying computation, communication, memory, and I/O intensity. A good understanding of the performance characteristics of each user application is important for job scheduling and resource provisioning. Among these performance characteristics, I/O performance is becoming increasingly important as data sizes rapidly increase and large-scale applications, such as simulation and model training, are widely adopted. However, predicting I/O performance is difficult because I/O systems are shared among all users and involve many layers of software and hardware stack, including the application, network interconnect, operating system, file system, and storage devices. Furthermore, updates to these layers and changes in system management policy can significantly alter the I/O behavior of applications and the entire system. To improve the prediction of the I/O performance on HPC systems, we propose integrating information from several different system logs and developing a regression-based approach to predict the I/O performance. Our proposed scheme can dynamically select the most relevant features from the log entries using various feature selection algorithms and scoring functions, and can automatically select the regression algorithm with the best accuracy for the prediction task. The evaluation results show that our proposed scheme can predict the write performance with up to 90% prediction accuracy and the read performance with up to 99% prediction accuracy using the real logs from the Cori supercomputer system at NERSC. [ABSTRACT FROM AUTHOR]

In this work, we have analyzed the input/output (I/O) activities of Cori, which is a high-performance computing system at the National Energy Research Scientific Computing Center at Lawrence Berkeley National Laboratory. Our analysis results indicate that most users do not adjust storage configurations but rather use the default settings. In addition, owing to the interference from many applications running simultaneously, the performance varies based on the system status. To configure file systems autonomously in complex environments, we developed DCA-IO, a dynamic distributed file system configuration adjustment algorithm that utilizes the system log information to adjust storage configurations automatically. Our scheme aims to improve the application performance and avoid interference from other applications without user intervention. Moreover, DCA-IO uses the existing system logs and does not require code modifications, an additional library, or user intervention. To demonstrate the effectiveness of DCA-IO, we performed experiments using I/O kernels of real applications in both an isolated small-sized Lustre environment and Cori. Our experimental results shows that our scheme can improve the performance of HPC applications by up to 263% with the default Lustre configuration. [ABSTRACT FROM AUTHOR]

Traffic analysis is a core element in network operations and management for various purposes including change detection, traffic prediction, and anomaly detection. In this paper, we introduce a new approach to online traffic analysis based on a pattern-based representation for high-level summarization of the traffic measurement data. Unlike the past online analysis techniques limited to a single variable to summarize (e.g., sketch), the focus of this study is on capturing the network state from the multivariate attributes under consideration. To this end, we employ clustering with its benefit of the aggregation of multidimensional variables. The clustered result represents the state of the network with regard to the monitored variables, which can also be compared with the observed patterns from previous time windows enabling intuitive analysis. We demonstrate the proposed method with two popular use cases, one for estimating state changes and the other for identifying anomalous states, to confirm its feasibility. Our extensive experimental results with public traces and collected monitoring measurements from ESnet traffic traces show that our pattern-based approach is effective for multivariate analysis of online network traffic with visual and quantitative tools. [ABSTRACT FROM AUTHOR]

Network traffic analysis is one of the core functions in network monitoring for effective network operations and management. While online traffic analysis has been widely studied, it is still intensively challenging due to several reasons. One of the primary challenges is the heavy volume of traffic to analyze within a finite amount of time due to the increasing network bandwidth. Another important challenge for effective traffic analysis is to support multivariate functions of traffic variables to help administrators identify unexpected network events intuitively. To this end, we propose a new approach with the multivariate analysis that offers a high-level summary of the online network traffic. With this approach, the current state of the network will display patterns compiled from a set of traffic variables, and the detection problems in network monitoring (e.g., change detection and anomaly detection) can be reduced to a pattern identification and classification problem. In this paper, we introduce our preliminary work with clustered patterns for online, multivariate network traffic analysis with the challenges and limitations we observed. We then present a grid-based model that is designed to overcome the limitations of the clustered pattern-based technique. We will discuss the potential of the new model with respect to the technical challenges including streaming-based computation and robustness to outliers. [ABSTRACT FROM AUTHOR]

With the increasing number of geographically distributed scientific collaborations and the growing sizes of scientific data, it has become challenging for users to achieve the best possible network performance on a shared network. We have developed a model to forecast expected bandwidth utilization on high-bandwidth wide area networks. The forecast model can improve the efficiency of the resource utilization and scheduling of data movements on high-bandwidth networks to accommodate ever increasing data volume for large-scale scientific data applications. A univariate time-series forecast model is developed with the Seasonal decomposition of Time series by Loess (STL) and the AutoRegressive Integrated Moving Average (ARIMA) on Simple Network Management Protocol (SNMP) path utilization measurement data. Compared with the traditional approach such as Box-Jenkins methodology to train the ARIMA model, our forecast model reduces computation time up to 92.6 %. It also shows resilience against abrupt network usage changes. Our forecast model conducts the large number of multi-step forecast, and the forecast errors are within the mean absolute deviation (MAD) of the monitored measurements. [ABSTRACT FROM AUTHOR]