arXiv subjects

Jianfeng Zhan

Publications and source records attributed to Jianfeng Zhan.

At least 91 records · Page 5Linked to original sources

arXiv2017

HybridTune: Spatio-temporal Data and Model Driven Performance Diagnosis for Big Data Systems

With tremendous growing interests in Big Data systems, analyzing and facilitating their performance improvement become increasingly important. Although there have much research efforts for improving Big Data systems performance, efficiently analysing and diagnosing performance bottlenecks over these massively distributed systems remain a major challenge. In this paper, we propose a spatio-temporal correlation analysis approach based on stage characteristic and distribution characteristic of Big Data applications, which can associate the multi-level performance data fine-grained. On the basis of correlation data, we define some priori rules, select features and vectorize the corresponding datasets for different performance bottlenecks, such as, workload imbalance, data skew, abnormal node and outlier metrics. And then, we utilize the data and model driven algorithms for bottlenecks detection and diagnosis. In addition, we design and develop a lightweight, extensible tool HybridTune, and validate the diagnosis effectiveness of our tool with BigDataBench on several benchmark experiments in which the outperform state-of-the-art methods. Our experiments show that the accuracy of abnormal/outlier detection we obtained reaches about 80%. At last, we report several Spark and Hadoop use cases, which are demonstrated how HybridTune supports users to carry out the performance analysis and diagnosis efficiently on the Spark and Hadoop applications, and our experiences demonstrate HybridTune can help users find the performance bottlenecks and provide optimization recommendations.

cs.DC↗

arXiv2017

A Dwarf-based Scalable Big Data Benchmarking Methodology

Different from the traditional benchmarking methodology that creates a new benchmark or proxy for every possible workload, this paper presents a scalable big data benchmarking methodology. Among a wide variety of big data analytics workloads, we identify eight big data dwarfs, each of which captures the common requirements of each class of unit of computation while being reasonably divorced from individual implementations. We implement the eight dwarfs on different software stacks, e.g., OpenMP, MPI, Hadoop as the dwarf components. For the purpose of architecture simulation, we construct and tune big data proxy benchmarks using the directed acyclic graph (DAG)-like combinations of the dwarf components with different weights to mimic the benchmarks in BigDataBench. Our proxy benchmarks preserve the micro-architecture, memory, and I/O characteristics, and they shorten the simulation time by 100s times while maintain the average micro-architectural data accuracy above 90 percentage on both X86 64 and ARMv8 processors. We will open-source the big data dwarf components and proxy benchmarks soon.

cs.AR↗

arXiv2017

Isolate First, Then Share: a New OS Architecture for Datacenter Computing

This paper presents the "isolate first, then share" OS model in which the processor cores, memory, and devices are divided up between disparate OS instances and a new abstraction, subOS, is proposed to encapsulate an OS instance that can be created, destroyed, and resized on-the-fly. The intuition is that this avoids shared kernel states between applications, which in turn reduces performance loss caused by contention. We decompose the OS into the supervisor and several subOSes running at the same privilege level: a subOS directly manages physical resources, while the supervisor can create, destroy, resize a subOS on-the-fly. The supervisor and subOSes have few state sharing, but fast inter-subOS communication mechanisms are provided on demand. We present the first implementation, RainForest, which supports unmodified Linux binaries. Our comprehensive evaluation shows RainForest outperforms Linux with four different kernels, LXC, and Xen in terms of worst-case and average performance most of time when running a large number of benchmarks. The source code is available soon.

cs.OS↗

arXiv2017

Cosine Normalization: Using Cosine Similarity Instead of Dot Product in Neural Networks

Traditionally, multi-layer neural networks use dot product between the output vector of previous layer and the incoming weight vector as the input to activation function. The result of dot product is unbounded, thus increases the risk of large variance. Large variance of neuron makes the model sensitive to the change of input distribution, thus results in poor generalization, and aggravates the internal covariate shift which slows down the training. To bound dot product and decrease the variance, we propose to use cosine similarity or centered cosine similarity (Pearson Correlation Coefficient) instead of dot product in neural networks, which we call cosine normalization. We compare cosine normalization with batch, weight and layer normalization in fully-connected neural networks as well as convolutional networks on the data sets of MNIST, 20NEWS GROUP, CIFAR-10/100 and SVHN. Experiments show that cosine normalization achieves better performance than other normalization techniques.

cs.LG↗

arXiv2017

10-millisecond Computing

Despite computation becomes much complex on data with an unprecedented scale, we argue computers or smart devices should and will consistently provide information and knowledge to human being in the order of a few tens milliseconds. We coin a new term 10-millisecond computing to call attention to this class of workloads. 10-millisecond computing raises many challenges for both software and hardware stacks. In this paper, using a typical workload-memcached on a 40-core server (a main-stream server in near future), we quantitatively measure 10-ms computing's challenges to conventional operating systems. For better communication, we propose a simple metric-outlier proportion to measure quality of service: for N completed requests or jobs, if M jobs or requests' latencies exceed the outlier threshold t, the outlier proportion is M/N . For a 1K-scale system running Linux (version 2.6.32), LXC (version 0.7.5) or XEN (version 4.0.0), respectively, we surprisingly find that so as to reduce the service outlier proportion to 10% (10% users will feel QoS degradation), the outlier proportion of a single server has to be reduced by 871X, 2372X, 2372X accordingly. Also, we discuss the possible design spaces of 10-ms computing systems from perspectives of datacenter architectures, networking, OS and scheduling, and benchmarking.

cs.PF↗

arXiv2016

AccuracyTrader: Accuracy-aware Approximate Processing for Low Tail Latency and High Result Accuracy in Cloud Online Services

Modern latency-critical online services such as search engines often process requests by consulting large input data spanning massive parallel components. Hence the tail latency of these components determines the service latency. To trade off result accuracy for tail latency reduction, existing techniques use the components responding before a specified deadline to produce approximate results. However, they may skip a large proportion of components when load gets heavier, thus incurring large accuracy losses. This paper presents AccuracyTrader that produces approximate results with small accuracy losses while maintaining low tail latency. AccuracyTrader aggregates information of input data on each component to create a small synopsis, thus enabling all components producing initial results quickly using their synopses. AccuracyTrader also uses synopses to identify the parts of input data most related to arbitrary requests' result accuracy, thus first using these parts to improve the produced results in order to minimize accuracy losses. We evaluated AccuracyTrader using workloads in real services. The results show: (i) AccuracyTrader reduces tail latency by over 40 times with accuracy losses of less than 7% compared to existing exact processing techniques; (ii) when using the same latency, AccuracyTrader reduces accuracy losses by over 13 times comparing to existing approximate processing techniques.

cs.DC↗

arXiv2015

PCS: Predictive Component-level Scheduling for Reducing Tail Latency in Cloud Online Services

Modern latency-critical online services often rely on composing results from a large number of server components. Hence the tail latency (e.g. the 99th percentile of response time), rather than the average, of these components determines the overall service performance. When hosted on a cloud environment, the components of a service typically co-locate with short batch jobs to increase machine utilizations, and share and contend resources such as caches and I/O bandwidths with them. The highly dynamic nature of batch jobs in terms of their workload types and input sizes causes continuously changing performance interference to individual components, hence leading to their latency variability and high tail latency. However, existing techniques either ignore such fine-grained component latency variability when managing service performance, or rely on executing redundant requests to reduce the tail latency, which adversely deteriorate the service performance when load gets heavier. In this paper, we propose PCS, a predictive and component-level scheduling framework to reduce tail latency for large-scale, parallel online services. It uses an analytical performance model to simultaneously predict the component latency and the overall service performance on different nodes. Based on the predicted performance, the scheduler identifies straggling components and conducts near-optimal component-node allocations to adapt to the changing performance interferences from batch jobs. We demonstrate that, using realistic workloads, the proposed scheduler reduces the component tail latency by an average of 67.05\% and the average overall service latency by 64.16\% compared with the state-of-the-art techniques on reducing tail latency.

cs.DC↗

arXiv2015

Characterization and Architectural Implications of Big Data Workloads

Big data areas are expanding in a fast way in terms of increasing workloads and runtime systems, and this situation imposes a serious challenge to workload characterization, which is the foundation of innovative system and architecture design. The previous major efforts on big data benchmarking either propose a comprehensive but a large amount of workloads, or only select a few workloads according to so-called popularity, which may lead to partial or even biased observations. In this paper, on the basis of a comprehensive big data benchmark suite---BigDataBench, we reduced 77 workloads to 17 representative workloads from a micro-architectural perspective. On a typical state-of-practice platform---Intel Xeon E5645, we compare the representative big data workloads with SPECINT, SPECCFP, PARSEC, CloudSuite and HPCC. After a comprehensive workload characterization, we have the following observations. First, the big data workloads are data movement dominated computing with more branch operations, taking up to 92% percentage in terms of instruction mix, which places them in a different class from Desktop (SPEC CPU2006), CMP (PARSEC), HPC (HPCC) workloads. Second, corroborating the previous work, Hadoop and Spark based big data workloads have higher front-end stalls. Comparing with the traditional workloads i. e. PARSEC, the big data workloads have larger instructions footprint. But we also note that, in addition to varied instruction-level parallelism, there are significant disparities of front-end efficiencies among different big data workloads. Third, we found complex software stacks that fail to use state-of-practise processors efficiently are one of the main factors leading to high front-end stalls. For the same workloads, the L1I cache miss rates have one order of magnitude differences among diverse implementations with different software stacks.

cs.DC↗

arXiv2015

Identifying Dwarfs Workloads in Big Data Analytics

Big data benchmarking is particularly important and provides applicable yardsticks for evaluating booming big data systems. However, wide coverage and great complexity of big data computing impose big challenges on big data benchmarking. How can we construct a benchmark suite using a minimum set of units of computation to represent diversity of big data analytics workloads? Big data dwarfs are abstractions of extracting frequently appearing operations in big data computing. One dwarf represents one unit of computation, and big data workloads are decomposed into one or more dwarfs. Furthermore, dwarfs workloads rather than vast real workloads are more cost-efficient and representative to evaluate big data systems. In this paper, we extensively investigate six most important or emerging application domains i.e. search engine, social network, e-commerce, multimedia, bioinformatics and astronomy. After analyzing forty representative algorithms, we single out eight dwarfs workloads in big data analytics other than OLAP, which are linear algebra, sampling, logic operations, transform operations, set operations, graph operations, statistic operations and sort.

cs.DB↗

arXiv2015

Understanding Big Data Analytic Workloads on Modern Processors

Big data analytics applications play a significant role in data centers, and hence it has become increasingly important to understand their behaviors in order to further improve the performance of data center computer systems, in which characterizing representative workloads is a key practical problem. In this paper, after investigating three most impor- tant application domains in terms of page views and daily visitors, we chose 11 repre- sentative data analytics workloads and characterized their micro-architectural behaviors by using hardware performance counters, so as to understand the impacts and implications of data analytics workloads on the systems equipped with modern superscalar out-of-order processors. Our study reveals that big data analytics applications themselves share many inherent characteristics, which place them in a different class from traditional workloads and scale-out services. To further understand the characteristics of big data analytics work- loads we performed a correlation analysis of CPI (cycles per instruction) with other micro- architecture level characteristics and an investigation of the big data software stack impacts on application behaviors. Our correlation analysis showed that even though big data ana- lytics workloads own notable pipeline front end stalls, the main factors affecting the CPI performance are long latency data accesses rather than the front end stalls. Our software stack investigation found that the typical big data software stack significantly contributes to the front end stalls and incurs bigger working set. Finally we gave several recommen- dations for architects, programmers and big data system designers with the knowledge acquired from this paper.

cs.DC↗

arXiv2014

Characterizing and Subsetting Big Data Workloads

Big data benchmark suites must include a diversity of data and workloads to be useful in fairly evaluating big data systems and architectures. However, using truly comprehensive benchmarks poses great challenges for the architecture community. First, we need to thoroughly understand the behaviors of a variety of workloads. Second, our usual simulation-based research methods become prohibitively expensive for big data. As big data is an emerging field, more and more software stacks are being proposed to facilitate the development of big data applications, which aggravates hese challenges. In this paper, we first use Principle Component Analysis (PCA) to identify the most important characteristics from 45 metrics to characterize big data workloads from BigDataBench, a comprehensive big data benchmark suite. Second, we apply a clustering technique to the principle components obtained from the PCA to investigate the similarity among big data workloads, and we verify the importance of including different software stacks for big data benchmarking. Third, we select seven representative big data workloads by removing redundant ones and release the BigDataBench simulation version, which is publicly available from http://prof.ict.ac.cn/BigDataBench/simulatorversion/.

cs.PF↗

arXiv2014

BDGS: A Scalable Big Data Generator Suite in Big Data Benchmarking

Data generation is a key issue in big data benchmarking that aims to generate application-specific data sets to meet the 4V requirements of big data. Specifically, big data generators need to generate scalable data (Volume) of different types (Variety) under controllable generation rates (Velocity) while keeping the important characteristics of raw data (Veracity). This gives rise to various new challenges about how we design generators efficiently and successfully. To date, most existing techniques can only generate limited types of data and support specific big data systems such as Hadoop. Hence we develop a tool, called Big Data Generator Suite (BDGS), to efficiently generate scalable big data while employing data models derived from real data to preserve data veracity. The effectiveness of BDGS is demonstrated by developing six data generators covering three representative data types (structured, semi-structured and unstructured) and three data sources (text, graph, and table data).

cs.DB↗

arXiv2014

BigDataBench: a Big Data Benchmark Suite from Internet Services

As architecture, systems, and data management communities pay greater attention to innovative big data systems and architectures, the pressure of benchmarking and evaluating these systems rises. Considering the broad use of big data systems, big data benchmarks must include diversity of data and workloads. Most of the state-of-the-art big data benchmarking efforts target evaluating specific types of applications or system software stacks, and hence they are not qualified for serving the purposes mentioned above. This paper presents our joint research efforts on this issue with several industrial partners. Our big data benchmark suite BigDataBench not only covers broad application scenarios, but also includes diverse and representative data sets. BigDataBench is publicly available from http://prof.ict.ac.cn/BigDataBench . Also, we comprehensively characterize 19 big data workloads included in BigDataBench with varying data inputs. On a typical state-of-practice processor, Intel Xeon E5645, we have the following observations: First, in comparison with the traditional benchmarks: including PARSEC, HPCC, and SPECCPU, big data applications have very low operation intensity; Second, the volume of data input has non-negligible impact on micro-architecture characteristics, which may impose challenges for simulation-based big data architecture research; Last but not least, corroborating the observations in CloudSuite and DCBench (which use smaller data inputs), we find that the numbers of L1 instruction cache misses per 1000 instructions of the big data applications are higher than in the traditional benchmarks; also, we find that L3 caches are effective for the big data applications, corroborating the observation in DCBench.

cs.DB↗

arXiv2013

The Implications of Diverse Applications and Scalable Data Sets in Benchmarking Big Data Systems

Now we live in an era of big data, and big data applications are becoming more and more pervasive. How to benchmark data center computer systems running big data applications (in short big data systems) is a hot topic. In this paper, we focus on measuring the performance impacts of diverse applications and scalable volumes of data sets on big data systems. For four typical data analysis applications---an important class of big data applications, we find two major results through experiments: first, the data scale has a significant impact on the performance of big data systems, so we must provide scalable volumes of data sets in big data benchmarks. Second, for the four applications, even all of them use the simple algorithms, the performance trends are different with increasing data scales, and hence we must consider not only variety of data sets but also variety of applications in benchmarking big data systems.

cs.PF↗

arXiv2013

Characterizing Data Analysis Workloads in Data Centers

As the amount of data explodes rapidly, more and more corporations are using data centers to make effective decisions and gain a competitive edge. Data analysis applications play a significant role in data centers, and hence it has became increasingly important to understand their behaviors in order to further improve the performance of data center computer systems. In this paper, after investigating three most important application domains in terms of page views and daily visitors, we choose eleven representative data analysis workloads and characterize their micro-architectural characteristics by using hardware performance counters, in order to understand the impacts and implications of data analysis workloads on the systems equipped with modern superscalar out-of-order processors. Our study on the workloads reveals that data analysis applications share many inherent characteristics, which place them in a different class from desktop (SPEC CPU2006), HPC (HPCC), and service workloads, including traditional server workloads (SPECweb2005) and scale-out service workloads (four among six benchmarks in CloudSuite), and accordingly we give several recommendations for architecture and system optimizations. On the basis of our workload characterization work, we released a benchmark suite named DCBench for typical datacenter workloads, including data analysis and service workloads, with an open-source license on our project home page on http://prof.ict.ac.cn/DCBench. We hope that DCBench is helpful for performing architecture and small-to-medium scale system researches for datacenter computing.

cs.PF↗

arXiv2013

BigDataBench: a Big Data Benchmark Suite from Web Search Engines

This paper presents our joint research efforts on big data benchmarking with several industrial partners. Considering the complexity, diversity, workload churns, and rapid evolution of big data systems, we take an incremental approach in big data benchmarking. For the first step, we pay attention to search engines, which are the most important domain in Internet services in terms of the number of page views and daily visitors. However, search engine service providers treat data, applications, and web access logs as business confidentiality, which prevents us from building benchmarks. To overcome those difficulties, with several industry partners, we widely investigated the open source solutions in search engines, and obtained the permission of using anonymous Web access logs. Moreover, with two years' great efforts, we created a sematic search engine named ProfSearch (available from http://prof.ict.ac.cn). These efforts pave the path for our big data benchmark suite from search engines---BigDataBench, which is released on the web page (http://prof.ict.ac.cn/BigDataBench). We report our detailed analysis of search engine workloads, and present our benchmarking methodology. An innovative data generation methodology and tool are proposed to generate scalable volumes of big data from a small seed of real data, preserving semantics and locality of data. Also, we preliminarily report two case studies using BigDataBench for both system and architecture researches.

cs.IR↗

arXiv2013

LogMaster: Mining Event Correlations in Logs of Large scale Cluster Systems

This paper presents a methodology and a system, named LogMaster, for mining correlations of events that have multiple attributions, i.e., node ID, application ID, event type, and event severity, in logs of large-scale cluster systems. Different from traditional transactional data, e.g., supermarket purchases, system logs have their unique characteristic, and hence we propose several innovative approaches to mine their correlations. We present a simple metrics to measure correlations of events that may happen interleavedly. On the basis of the measurement of correlations, we propose two approaches to mine event correlations; meanwhile, we propose an innovative abstraction: event correlation graphs (ECGs) to represent event correlations, and present an ECGs based algorithm for predicting events. For two system logs of a production Hadoop-based cloud computing system at Research Institution of China Mobile and a production HPC cluster system at Los Alamos National Lab (LANL), we evaluate our approaches in three scenarios: (a) predicting all events on the basis of both failure and non-failure events; (b) predicting only failure events on the basis of both failure and non-failure events; (c) predicting failure events after removing non-failure events.

cs.DC↗

arXiv2013

High Volume Computing: Identifying and Characterizing Throughput Oriented Workloads in Data Centers

For the first time, this paper systematically identifies three categories of throughput oriented workloads in data centers: services, data processing applications, and interactive real-time applications, whose targets are to increase the volume of throughput in terms of processed requests or data, or supported maximum number of simultaneous subscribers, respectively, and we coin a new term high volume computing (in short HVC) to describe those workloads and data center computer systems designed for them. We characterize and compare HVC with other computing paradigms, e.g., high throughput computing, warehouse-scale computing, and cloud computing, in terms of levels, workloads, metrics, coupling degree, data scales, and number of jobs or service instances. We also preliminarily report our ongoing work on the metrics and benchmarks for HVC systems, which is the foundation of designing innovative data center computer systems for HVC workloads.

cs.DC↗