Data Transfer Size Research Articles

Modeling inter-DC bulk data transfer in optical networks with edge storage

In communication networks that span large geographical regions, transfer of bulk data is often costly and hard to manage. One prominent reason for that is bandwidth contention among long-lasting flows, and between long-lasting and short-lived ones on inter-datacenter links. Delaying the transfer of bulk data to off-peak hours by temporarily storing transfer at edge of the network can help reduce bandwidth contention during peak hours, and improve overall bandwidth utilization and network utility. Although extensive research has been conducted on its use, to date no model has been developed to quantify the performance of storage assisted bulk transfers across multiple congested links. This research gap has limited our understanding of the operation of storage assisted transfer under different network conditions. In this paper, we model bulk data transfer in inter-datacenter optical networks with edge storage, and derive closed-form equations for bulk data transfer on inter-DC links that exhibit diurnal traffic patterns and varying peak hours. Our study reveals that when the data transfer requests arrive uniformly in a day, the probability of a successful transfer increases linearly with the time data is allowed to wait in edge storage. Our study also shows that this probability decreases linearly with β + D, where β and D are the durations of background traffic peak and size of data transfers, respectively, when the network is moderately loaded. And, the benefit of storage will barely increase when the allowed waiting time goes beyond β + D.

The Bitlet Model: A Parameterized Analytical Model to Compare PIM and CPU Systems

Currently, data-intensive applications are gaining popularity. Together with this trend, processing-in-memory (PIM)–based systems are being given more attention and have become more relevant. This article describes an analytical modeling tool called Bitlet that can be used in a parameterized fashion to estimate the performance and power/energy of a PIM-based system and, thereby, assess the affinity of workloads for PIM as opposed to traditional computing. The tool uncovers interesting trade-offs between, mainly, the PIM computation complexity (cycles required to perform a computation through PIM), the amount of memory used for PIM, the system memory bandwidth, and the data transfer size. Despite its simplicity, the model reveals new insights when applied to real-life examples. The model is demonstrated for several synthetic examples and then applied to explore the influence of different parameters on two systems — IMAGING and FloatPIM. Based on the demonstrations, insights about PIM and its combination with a CPU are provided.

Enhanced Clients for Data Stores and Cloud Services

Data stores and cloud services are typically accessed using a client-server paradigm wherein the client runs as part of an application process which is trying to access the data store or cloud service. This paper presents the design and implementation of enhanced clients for improving both the functionality and performance of applications accessing data stores or cloud services. Our enhanced clients can improve performance via multiple types of caches, encrypt data for providing confidentiality before sending information to a server, and compress data for reducing the size of data transfers. Our clients can perform data analysis to allow applications to more effectively use cloud services. They also provide both synchronous and asynchronous interfaces. An asynchronous interface allows an application program to access a data store or cloud service and continue execution before receiving a response which can significantly improve performance. We present a Universal Data Store Manager (UDSM) which allows an application to access multiple different data stores and provides a common interface to each data store. The UDSM also can monitor the performance of different data stores. A workload generator allows users to easily determine and compare the performance of different data stores. We also present NLU-SA, an application for performing natural language understanding and sentiment analysis on text documents. NLU-SA is implemented on top of our enhanced clients and integrates text analysis with Web searching. We present results from NLU-SA on sentiment on the Web towards major companies and countries. We also present a performance analysis of our enhanced clients.

Energy Efficiency of Image Transmission in Embedded Linux based Wireless Visual Sensor Network

Wireless Visual Sensor Network (WVSN) is a system that consists of visual sensor nodes with an embedded processor. WVSN devices have limited resources of energy, computation capability, memory, and bandwidth. Due to these limitations the implementation of WVSN for large multimedia data, such as images, become a challenging task. Therefore, it is required compressed images prior to transmission. In addition to the limited resources, the system implementation strongly affects the efficiency of the working system.The main contribution of this research is to offer a technical solution of simpler image compression on the WVSN platform. JPEG 2000 is investigated as an alternative compression method to reduce the size of data transfer on WVSN using Embedded Linux as its operating system. Compressed images are transferred to a receiver on communication of IEEE 802.15.4.. This paper shows that the energy consumption for compression and transmission will reduce to only 10.48%, 13.60%, and 17.11% compared to raw image. BER will significantly reduce by implementing image compression. Therefore, it is demonstrated that this model significantly increases energy efficiency, memory utilization efficiency, and data transfer time with acceptable PSNR, compared to uncompressed images.

A regression‐based performance prediction framework for synchronous iterative algorithms on general purpose graphical processing unit clusters

SUMMARYHeterogeneous performance prediction models are valuable tools to accurately predict application runtime, allowing for efficient design space exploration and application mapping. The existing performance models require intricate system architecture knowledge, making the modeling task difficult. In this research, we propose a regression‐based performance prediction framework for general purpose graphical processing unit (GPGPU) clusters that statistically abstracts the system architecture characteristics, enabling performance prediction without detailed system architecture knowledge. The regression‐based framework targets deterministic synchronous iterative algorithms using our synchronous iterative GPGPU execution model and is broken into two components: the computation component that models the GPGPU device and host computations and the communication component that models the network‐level communications. The computation component regression models use algorithm characteristics such as the number of floating‐point operations and total bytes as predictor variables and are trained using several small, instrumented executions of synchronous iterative algorithms that include a range of floating‐point operations‐to‐byte requirements. The regression models for network‐level communications are developed using micro‐benchmarks and employ data transfer size and processor count as predictor variables. Our performance prediction framework achieves prediction accuracy over 90% compared with the actual implementations for several tested GPGPU cluster configurations. The end goal of this research is to offer the scientific computing community, an accurate and easy‐to‐use performance prediction framework that empowers users to optimally utilize the heterogeneous resources. Copyright © 2013 John Wiley & Sons, Ltd.

Characteristics of Forensic Imaging Performance—An Analysis of Forensic Imaging Bottlenecks

Disk imaging involves copying all of the data from a source disk drive to a target. Typically, the target for the copy is another disk drive. Forensic processes developed years ago do not appear to be adequate for current storage technology. For example, with disk drive capacities now exceeding 1 Terabyte, a typical disk imaging can take over 8 hours at typical rates. With disk drive capacities increasing, forensic copying is expected to take even longer. Along with increase in disk capacity, the industry has also seen an increase in data transfer rates. In many cases, forensic imaging is taking longer than necessary. To identify the bottlenecks, an examination of different methods used to transfer data from a source disk was performed. Factors considered were differing disk access technologies. One finding is that the USB disk access technology (version 2.0 and earlier) is a significant bottleneck for data transfer rates, especially when the USB device is a write-blocker. Other factors that contribute to the efficiency of a forensic copy are the file system used to write a forensic image and the data transfer size used when reading from a disk drive. Optimal parameters for performing a forensic acquisition from a disk drive are identified.

Reducing On-Chip DRAM Energy via Data Transfer Size Optimization

This paper proposes a software-controllable variable line-size (SC-VLS) cache architecture for low power embedded systems. High bandwidth between logic and a DRAM is realized by means of advanced integrated technology. System-in-Silicon is one of the architectural frameworks to realize the high bandwidth. An ASIC and a specific SRAM are mounted onto a silicon interposer. Each chip is connected to the silicon interposer by eutectic solder bumps. In the framework, it is important to reduce the DRAM energy consumption. The specific DRAM needs a small cache memory to improve the performance. We exploit the cache to reduce the DRAM energy consumption. During application program executions, an adequate cache line size which produces the lowest cache miss ratio is varied because the amount of spatial locality of memory references changes. If we employ a large cache line size, we can expect the effect of prefetching. However, the DRAM energy consumption is larger than a small line size because of the huge number of banks are accessed. The SC-VLS cache is able to change a line size to an adequate one at runtime with a small area and power overheads. We analyze the adequate line size and insert line size change instructions at the beginning of each function of a target program before executing the program. In our evaluation, it is observed that the SC-VLS cache reduces the DRAM energy consumption up to 88%, compared to a conventional cache with fixed 256 B lines.

Routing and scheduling connections in networks that support advance reservations

A key problem in networks that support advance reservations is the routing and time scheduling of connections with flexible starting time and known data transfer size. In this paper we present a multicost routing and scheduling algorithm for selecting the path to be followed by such a connection and the time the data should start and end transmission at each link so as to minimize the reception time at the destination, or optimize some other performance criterion. The utilization profiles of the network links, the link propagation delays, and the parameters of the connection to be scheduled form the inputs to the algorithm. We initially present a scheme of non-polynomial complexity to compute a set of so-called non-dominated candidate paths, from which the optimal path can be found. We then propose two mechanisms to appropriately prune the set of candidate paths in order to find multicost routing and scheduling algorithms of polynomial complexity. We examine the performance of the algorithms in the special case of an Optical Burst Switched network. Our results indicate that the proposed polynomial-time algorithms have performance that is very close to that of the optimal algorithm. We also study the effects network propagation delays and link-state update policies have on performance.

On the effectiveness of movement prediction to reduce energy consumption in wireless communication

Node movement can be exploited to reduce the energy consumption of wireless network communication. The strategy consists in delaying communication until a mobile node moves close to its target peer node within an application- imposed deadline. We evaluate the performance of various heuristics that, based on the movement history of the mobile node, estimate an optimal time (in the sense of least energy use) of communication subject to the delay constraint. We evaluate the impact of the node movement model, length of movement history maintained, allowable delay, single hop versus multiple hop communication, and size of data transfer on the energy consumption. We also present measurement results on an iPAQ pocket PC that quantity energy consumption in executing the prediction algorithms. Our results show that, with relatively simple and, hence, efficient prediction heuristics, energy savings in communication can significantly outweigh the energy expenses in executing the prediction algorithms. Moreover, it is possible to achieve robust system performance across diverse node movement models.

On the effectiveness of movement prediction to reduce energy consumption in wireless communication

Node movement can be exploited to reduce the energy consumption of wireless network communication. The strategy consists in delaying communication until a mobile node moves close to its target peer node, within an application-imposed deadline. We evaluate the performance of various heuristics that, based on the movement history of the mobile node, estimate an optimal time (in the sense of least energy use) of communication subject to the delay constraint. We evaluate the impact of node movement model, length of movement history maintained, allowable delay, single hop versus multiple hop communication, and size of data transfer on the energy consumption. We also present measurement results on an iPAQ pocket PC that quantify energy consumption in executing the prediction algorithms. Our results show that, with relatively simple and hence efficient prediction heuristics, energy savings in communication can significantly outweigh the energy expenses in executing the prediction algorithms.

데이터베이스 클러스터의 가용성 향상을 위한 온라인 확장 기법

온라인 확장 기법은 비공유 데이터베이스 클러스터에서 온라인 상태에서 새로운 노드를 추가하고 데이터 재조직을 수행함으로써 작업 부하를 분산시키거나 전체 트랜잭션 처리량을 늘리기 위한 기법이다. 그러나, 기존의 온라인 확장 기법에서는 과부하 상태의 노드에 데이터 전송과 일관성 유지에 대한 추가적인 부하가 발생됨으로써, 전체 시스템의 응답속도가 느려지고 노드의 결함 발생 가능성을 감소시키는 문제점이 있다. 본 논문에서는 비공유 데이터베이스 클러스터에서 높은 가용성을 위한 데이터 확장 기법으 제안한다. 제안딘 온라인 확장 기법은 확장 연산 수행 중에 발생되는 노드의 추가적 부하를 병렬 데이터 전송과정과 복제본의 완성 과정을 통해 분산시키고, 확장 중에 발생한 결함에 대해서 효율적인 회복을 수행함으로써 데이터베이스 클러스터의 가용성을 향상시킨다. 즉, 원본 노드의 데이터를 각 복제분이 저장된 노드들에서 동시에 전송함으로써 데이터 전송을 병렬화하고, 전송 영역을 서로 분배하여 원본 노드의 데이터을 각 복제본이 저장된 노드들에서 동시에 전송함으로써 데이터 전송을 병렬화사고, 전송영역ㅇ 버로 분배하여 원본 노드의 부하와 다른 트랜잭선에 대한 간섭을 줄인다. 또한, 온라인 확장 기법에서의 노드 결함에 대해 빠른 회복을 수행한다. 본 논문에서는 성능평가를 통해 제안 기법이 기존 기법에 비해 노드의 부하를 감소시켜 결함 발생 가능성을 낮추고, 온라인 확장 연산에 대한 회복 처리 시간을 단축하여 데이터베이스 클러스터의 가용성을 향상시킴을 보인다. An online scaling method adds new nodes to the shared-nothing database cluster and makes tables be reorganized while the system is running. The objective is to share the workload with many nodes and increase the capacity of cluster systems. The existing online scaling method, however, has two problems. One is the degradation of response time and transactions throughput due to the additional overheads of data transfer and replica's condidtency. The other is and inefficient recovery mechanism in which the overall scaling transaction is aborted by a fault. These problems deteriorate the availability of shared-nothing database cluster. To avoid the additional overheads throughout the scaling period, our scalingmethod consists of twophases : a parallel data transfer phase and a combination phase. The parallel data transferred datausing reduces the size of data transfer by dividing the data into the number of replicas. The combination phase combines the transferred datausing resources of spare nodes. Also, our method reduces the possibility of failure throughout the scaling period and improves the availability of the database cluster.

Improving index performance through prefetching

This paper proposes and evaluate Prefetching B + - Trees (pB + -Trees), which use prefetching to accelerate two important operations on B + -Tree indices: searches and range scans. To accelerate searches, pB + -Trees use prefetching to effectively create wider nodes than the natural data transfer size: e.g., eight vs. one cache lines or disk pages. These wider nodes reduce the height of the B + -Tree, thereby decreasing the number of expensive misses when going from parent to child without significantly increasing the cost of fetching a given node. Our results show that this technique speeds up search and update times by a factor of 1.21-1.5 for main-memory B + -Trees. In addition, it outperforms and is complementary to “Cache-Sensitive B + -Trees.” To accelerate range scans, pB + -Trees provide arrays of pointers to their leaf nodes. These allow the pB + -Tree to prefetch arbitrarily far ahead, even for nonclustered indices, thereby hiding the normally expensive cache misses associated with traversing the leaves within the range. Our results show that this technique yields over a sixfold speedup on range scans of 1000+ keys. Although our experimental evaluation focuses on main memory databases, the techniques that we propose are also applicable to hiding disk latency.

A new hierarchical disk architecture

This three-level hierarchical architecture exploits the temporal locality of disk accesses and the benefits of large data transfer sizes to improve disk drive write performance.