不确定数据流上的并行Skyline查询算法

不确定数据流上的Skyline查询技术逐步引起研究者的关注,传统的集中式流处理算法难以满足海量数据的查询需求,并且云计算所提供的海量计算资源和有效的存储管理模式,为研究并行Skyline查询技术提供了充足的条件。基于上述事实,提出了一种不确定数据流上的并行Skyline查询算法(parallel Skyline over uncertain data streams,PSUDS)。该算法通过交叉划分滑动窗口的方式,将集中式流查询转化为并行处理,以并行执行的方式来解决集中式算法处理性能不足的问题。大量实验结果表明,该算法具有较好的并行可扩展性。     

Online anomaly detection for multi‐source VMware using a distributed streaming framework

Anomaly detection refers to the identification of patterns in a dataset that do not conform to expected patterns. Such non‐conformant patterns typically correspond to samples of interest and are assigned to different labels in different domains, such as outliers, anomalies, exceptions, and malware. A daunting challenge is to detect anomalies in rapid voluminous streams of data. This paper presents a novel, generic real‐time distributed anomaly detection framework for multi‐source stream data. As a case study, we investigate anomaly detection for a multi‐source VMware‐based cloud data center, which maintains a large number of virtual machines (VMs). This framework continuously monitors VMware performance stream data related to CPU statistics (e.g., load and usage). It collects data simultaneously from all of the VMs connected to the network and notifies the resource manager to reschedule its CPU resources dynamically when it identifies any abnormal behavior from its collected data. A semi‐supervised clustering technique is used to build a model from benign training data only. During testing, if a data instance deviates significantly from the model, then it is flagged as an anomaly. Effective anomaly detection in this case demands a distributed framework with high throughput and low latency. Distributed streaming frameworks like Apache Storm, Apache Spark, S4, and others are designed for a lower data processing time and a higher throughput than standard centralized frameworks. We have experimentally compared the average processing latency of a tuple during clustering and prediction in both Spark and Storm and demonstrated that Spark processes a tuple much quicker than storm on average. Copyright © 2016 John Wiley & Sons, Ltd.     

基于Apache Flink的RDF流数据查询

目前成熟的RDF流处理（RDF Stream Processing, RSP）系统由于集中式的设计而缺乏并行处理特性,因此在查询处理大量传入的RDF流数据时,均无法实现高吞吐和低延迟。为提高查询性能,本文对RSP查询过程和Flink流计算结构进行研究,设计数据源、滤器、多路分区连接和投影4个逻辑操作符,并设计一种多流连接（Multi-Stream Join, MSJ）算法用于生成具有并行性的有向无环图的逻辑查询计划,最后以大数据流处理平台Apache Flink为底层实现逻辑操作符和逻辑查询计划。使用真实数据集SRBench和模拟数据集LUBMs进行实验验证。结果表明,与最成熟的系统C-SPARQL、CQELS相比,单机吞吐量增长高达10倍,5台机器集群的吞吐量增长高达28倍,同时在延时方面达到了毫秒级;在查询性能方面实现了处理大量RDF流数据时吞吐量的提高和延时的降低。     

高速连续数据流记录系统中并行处理接口的研究

为了解决高速数据流的连续记录/读取与存储介质速度慢之间的矛盾,文中用FPGA设计了基于RAID结构的并行处理接口,实现了高速数据的分割降速、合并/恢复、纠错重构,解决了高速数据流连续存储中的I/O瓶颈问题。并行处理接口采用了流水线的设计方式及动态的逻辑配置,使得系统性能得到很大的优化,解决了高速数据处理中的延迟、数据错误、工作时序不同步等问题。并行处理接口最终在实验系统中实现了对高达160MB/S连续实时数据流的处理。     

非规则流中高维数据流典型相关性分析并行计算方法

为了满足在计算资源受限的环境下高维数据流处理的实时性要求,提出一种方法——基于GPU(graphic processing unit)的非规则流中高维数据流的处理模型和具体的可行架构,并分析设计了相关的并行算法.该六层模型是将GPU处理数据的高宽带性能结合进滑动窗口中数据流的分析,进而在该框架下基于统一计算设备架构(compute unified device architecture,简称CUDA),使用数据立方模型以及降维约简技术并行分析了多条高维数据流的典型相关性.理论分析和实验结果均表明,该并行处理方法能够在线精确地识别同步滑动窗口模式下高维数据流之间的相关性.相对于纯CPU方法,该方法具有显著的速度优势,很好地满足了高维数据流的实时性需求,可以作为通用的分析方法广泛应用于数据流挖掘领域.     

Exploiting punctuation semantics in continuous data streams

As most current query processing architectures are already pipelined, it seems logical to apply them to data streams. However, two classes of query operators are impractical for processing long or infinite data streams. Unbounded stateful operators maintain state with no upper bound in size and, so, run out of memory. Blocking operators read an entire input before emitting a single output and, so, might never produce a result. We believe that a priori knowledge of a data stream can permit the use of such operators in some cases. We discuss a kind of stream semantics called punctuated streams. Punctuations in a stream mark the end of substreams allowing us to view an infinite stream as a mixture of finite streams. We introduce three kinds of invariants to specify the proper behavior of operators in the presence of punctuation. Pass invariants define when results can be passed on. Keep invariants define what must be kept in local state to continue successful operation. Propagation invariants define when punctuation can be passed on. We report on our initial implementation and show a strategy for proving implementations of these invariants are faithful to their relational counterparts.     

Enhancing throughput for streaming applications running on cluster systems

The exploitation of throughput in a parallel application that processes an input data stream is a difficult challenge. For typical coarse-grain applications, where the computation time of tasks is greater than their communication time, the maximum achievable throughput is determined by the maximum task computation time. Thus, the improvement in throughput above this maximum would eventually require the modification of the source code of the tasks. In this work, we address the improvement of throughput by proposing two task replication methodologies that have the target throughput to be achieved as an input parameter. They proceed by generating a new task graph structure that permits the target throughput to be achieved. The first replication mechanism, named DPRM (Data Parallel Replication Mechanism), exploits the inner task data parallelism. The second mechanism, named TCRM (Task Copy Replication Mechanism), creates new execution paths inside the application task graph structure that allows more than one instance of data to be processed concurrently. We evaluate the effectiveness of these mechanisms with three real applications executed in a cluster system: the MPEG2 video compressor, the IVUS (Intra-Vascular Ultra-Sound) medical image application and the BASIZ (Bright and SAtured Images Zone) video processing application. In all these cases, the obtained throughput was greater after applying the proposed replication mechanism than what the application could provide with the original implementation.     

基于流式计算的空间科学卫星数据实时处理

针对空间科学卫星探测数据的实时处理要求越来越高的问题,提出一种基于流计算框架的空间科学卫星数据实时处理方法。首先,根据空间科学卫星数据处理特点对数据流进行抽象分析;然后,对各处理单元的输入输出数据结构进行重新定义;最后,基于流计算框架Storm设计数据流处理并行结构,以适应大规模数据并行处理和分布式计算的要求。对应用该方法开发的空间科学卫星数据处理系统进行测试分析,测试结果显示,在相同条件下数据处理时间比原有系统缩短了一半;数据局部性策略比轮询策略具有更高的吞吐率,数据元组吞吐率平均提高29%。可见采用流式计算框架能够大幅缩短数据处理延迟,提高空间科学卫星数据处理系统的实时性。     

基于图形处理器的通用计算模式*

针对GPU图形处理的特点,分析其应用于通用计算的并行处理机制和数据映射,提出了一种GPU通用计算模式的映射机制和一般性设计方法,并针对GPU的吞吐量、数据流处理能力和基本数学运算能力等进行性能测试,为GPU通用计算的算法设计、实现和性能优化提供参考依据。     

基于流式计算的遥感卫星数据快视处理方法

随着高分辨率遥感卫星数据获取能力和地面数传接收能力的提高，现有遥感卫星快视处理系统的处理负载增大，实时性要求越来越难以满足。针对这些问题，采用流式计算思想提出了一种新的遥感卫星数据快视处理系统设计方法。在分析遥感卫星数据快视处理数据流特点的基础上，应用Storm框架对现有系统进行并行优化，设计遥感数据流处理任务拓扑结构，同时利用消息队列中间件Kafka改进处理单元间数据交换和数据缓存方式。实验表明，该系统在数据吞吐率和可靠性方面测试效果良好。     

应对倾斜数据流在线连接方法

并行环境下的分布式连接处理要求制定划分策略以减少状态迁移和通信开销。相对于数据库管理系统而言,分布式数据流管理系统中的在线θ连接操作需要更高的计算成本和内存资源。基于完全二部图的连接模型可支持分布式数据流的连接操作。因为连接操作的每个关系仅存放于二部图模型的一侧处理单元,无需复制数据,且处理单元相互独立,因此该模型具有内存高效、易伸缩和可扩展等特性。然而,由于数据流速的不稳定性和属性值分布的不均衡性,导致倾斜数据流的连接操作易出现集群负载不均衡的现象。针对倾斜数据流的连接操作,模型无法动态分配查询节点,并需要人工干预数据分组的参数设置。尤其是应对全部历史数据的连接查询,模型效率更低。基于上述问题,提出了管理倾斜数据流连接的框架,使用基于键值和元组混合的划分样式有效应对二部图模型的各侧倾斜数据。并设计了重新动态分配查询节点的策略和状态迁移算法,以支持全历史数据的连接查询和自适应的资源管理。针对合成数据和真实数据的实验表明,该方案可有效应对倾斜数据的连接操作并进一步提升分布式数据流管理系统的吞吐率,特别是降低云环境中的计算成本。     

易变数据流的系统资源配置方法

大规模数据流管理系统往往由上层的关系查询系统和下层的流处理系统组成。当用户提交查询请求时,往往需要根据数据流的流速和分布情况动态配置系统参数。然而,由于数据流的易变性,频繁改变参数配置会降低系统性能。针对该问题,提出了OrientStream+框架。设定以用户自定义查询延迟阈值为间隔片段的微批量数据流传输机制;并利用多级别管道缓存,对相同配置的数据流进行批量处理;然后按照数据流的时间戳计算出精准查询结果;引入基于异常检测的增量学习模型,用于提高OrientStream+的预测精度。最后,在Storm上实现了该资源配置框架,并进行了大量的实验。实验结果表明,OrientStream+框架可进一步降低系统的处理延迟并提高系统的吞吐率。     

A Multi-Domain Architecture for Mining Frequent Items and Itemsets from Distributed Data Streams

Real-time analysis of distributed data streams is a challenging task since it requires scalable solutions to handle streams of data that are generated very rapidly by multiple sources. This paper presents the design and the implementation of an architecture for the analysis of data streams in distributed environments. In particular, data stream analysis has been carried out for the computation of items and itemsets that exceed a frequency threshold. The mining approach is hybrid, that is, frequent items are calculated with a single pass, using a sketch algorithm, while frequent itemsets are calculated by a further multi-pass analysis. The architecture combines parallel and distributed processing to keep the pace with the rate of distributed data streams. In order to keep computation close to data, miners are distributed among the domains where data streams are generated. The paper reports the experimental results obtained with a prototype of the architecture, tested on a Grid composed of three domains each one handling a data stream.     

Distributed stream join under workload variance

Flexible and self-adaptive stream join processing plays an important role in a parallel shared-nothing environments. Join-Matrix model is a high-performance model which is resilient to data skew and supports arbitrary join predicates for taking random tuple distribution as its routing policy. To maximize system throughputs and minimize network communication cost, a scalable partitioning scheme on matrix is critical. In this paper, we present a novel flexible and adaptive scheme partitioning model for stream join operator, which ensures high throughput but with economical resource usages by allocating resources on demand. Specifically, a lightweight scheme generator, which requires the sample of each stream volume and processing resource quota of each physical machine, generates a join scheme; then a migration plan generator decides how to migrate data among machines under the consideration of minimizing migration cost while ensuring correctness. We do extensive experiments on different kinds of join workloads and the evaluation shows high competence comparing with baseline systems on benchmark data and real data.     

Parallel processing of continuous queries over data streams

In this paper, we propose parallel processing of continuous queries over data streams to handle the bottleneck of single processor DSMSs. Queries are executed in parallel over the logical machines in a multiprocessing environment. Scheduling parallel execution of operators is performed via finding the shortest path in a weighted graph called Query Mega Graph (QMG), which is a logical view of K machines. By lapse of time, number of tuples waiting in queues of different operators may be very different. When a queue becomes full, re-scheduling is done by updating weight of edges of QMG. In the new computed path, machines with more workload will be used less. The proposed system is formally presented and its correctness is proved. It is also modeled in PetriNets and its performance is evaluated and compared with serial query processing as well as the Min-Latency scheduling algorithm. The presented system is shown to outperform them w.r.t. tuple latency (response time), memory usage, throughput and also tuple loss- critical parameters in any data stream management systems.     

基于VLIW DSP 加密与认证算法的实现

针对HD视频数据流传输过程中数据安全与完整性问题,介绍了一种专用于DES、3DES、SHA1、MD5、RSA的加密VLIW DSP(LILY-DSP)。 为了提高性能和降低成本,DSP设计成具有11级流水线和2个并行执行簇,每一簇具有3个功能单元的专用并行结构。为了提高速率,定制了专用指令实现复杂操作。提出了基于此DSP的对称加密算法、公钥加密和认证算法实现方法。仿真实现结果表明,基于VLIW DSP加密和认证算法能很好地满足实时HD video数据流需求。     

A pipeline-based approach for scheduling video processing algorithms on NOW

Network Of Workstations (NOW) platforms put together with off-the-shelf workstations and networking hardware have become a cost effective, scalable, and flexible platform for video processing applications. Still, one has to manually schedule an algorithm to the available processors of the NOW to make efficient use of the resources. However, this approach is time-consuming and impractical for a video processing system that must perform a variety of different algorithms, with new algorithms being constantly developed. Improved support for program development is absolutely necessary before the full benefits of parallel architectures can be realized for video processing applications. Toward this goal, an automatic compile-time scheduler has been developed to schedule input tasks of video processing applications with precedence constraints onto available processors. The scheduler exploits both spatial (parallelism) and temporal (pipelining) concurrency to make the best use of machine resources. Two important scheduling problems are addressed. First, given a task graph and a desired throughput, a schedule is constructed to achieve the desired throughput with the minimum number of processors. Second, given a task graph and a finite set of available resources, a schedule is constructed such that the throughput is maximized while meeting the resource constraints. Results from simulations show that the scheduler and proposed optimization techniques effectively tackle these problems by maximizing processor utilization. A code generator has been developed to generate parallel programs automatically. The tools developed in this paper make it much easier for a programmer to develop video processing applications on these parallel architectures.     

使用GPU技术的数据流分位数并行计算方法

数据流实时、连续、快速到达的特点决定了数据流的实时处理能力。在处理低维数据流时经常使用分位数信息来描述数据流的统计信息,利用图形处理器（GPU）的强大计算能力和高内存带宽的特性计算数据流分位数信息,提出了基于统一计算设备架构（CUDA）的数据流处理模型和基于该模型的数据流分位数并行计算方法。实验证明,该方法在提供不低于纯CPU分位数算法相同精度的条件下,使数据流分位数的实时计算带宽得到了显著的提高。     

基于GPU的多数据流相关系数并行计算方法研究*

为了满足多数据流处理的实时性需求,提出一种跨PCIE总线的四层滑动窗口模型和基于图形处理器的多数据流并行处理框架模型,在此框架模型下可以并行维护数量巨大的滑动实时多数据流统计信息,同时采用精确方法并行计算多数据流间任意两条的相关系数。通过对比在同样的实验环境下只使用CPU的计算处理方法,验证了新方法的实时计算性能具有显著的提高。     

Functional programming with streams —Part II—

This paper complements our previous paper “Functional programming with streams.” The purpose of this paper is two-fold: to further develop the concept of a stream, and to present an implementation aspect of stream programming. Stream programming is decomposed into three phases, i.e. stream generation, stream transformation and stream reduction and for each phase we have (stream) generators, transformers and reducers, respectively. A linear recursive function equation, for example, is described as a composition of a stream generator and a reducer. We also give a listing of implemented stream processing functions in this paper.