With the rise of the robot and cloud computing technology, human-centric healthcare service attracts widely attention in order to meet the great challenges of traditional healthcare, especially the limited medical resources. This paper presents a healthcare system based on cloud computing and robotics, which consists of wireless body area networks, robot, software system and cloud platform. This system is expected to accurately measure user's physiological information for analysis and feedback, which is assisted by the robot integrated with various sensors. In order to improve the practicability of multi-media transmission in the healthcare system, this paper proposes a novel scheme to deliver real-time video contents through an improved UDP-based protocol. Finally, the proposed approach is evaluated with experimental study based on a real testbed.

Robot and cloud-assisted multi-modal healthcare system

Hadoop MapReduce Performance on SSDs for Analyzing Social Networks

A Survey on Data-driven Performance Tuning for Big Data Analytics Platforms

Given the rapid growth in cloud computing, it is important to analyze the performance of different Hadoop MapReduce applications and to understand the performance bottleneck in a cloud cluster that contributes to higher or lower performance. It is also important to analyze the underlying hardware in cloud cluster servers to enable the optimization of software and hardware to achieve the maximum performance possible. Hadoop is based on MapReduce, which is one of the most popular programming models for big data analysis in a parallel computing environment. In this paper, we present a detailed performance analysis, characterization, and evaluation of Hadoop MapReduce WordCount application. We also propose an estimation model based on Amdahl’s law regression method to estimate performance and total processing time versus different input sizes for a given processor architecture. The estimation regression model is verified to estimate performance and run time with an error margin of <5%.

Performance Evaluation and Estimation Model Using Regression Method for Hadoop WordCount

Scientists are increasingly using the current state of the art big data analytic software (e.g., Hadoop, Giraph, etc.) for their data-intensive applications over HPC environment. However, understanding and designing the hardware environment that these data- and compute-intensive applications require for good performance is challenging. With this motivation, we evaluated the performance of big data software over three different distributed-cyber-infrastructures, including a traditional HPC-cluster called SuperMikeII, a regular datacenter called SwatIII, and a novel MicroBrick-based hyperscale system called CeresII, using our own benchmark Parallel Genome Assembler (PGA). PGA is developed atop Hadoop and Giraph and serves as a good real-world example of a data- as well as compute-intensive workload. To evaluate the impact of both individual hardware components as well as overall organization, we changed the configuration of SwatIII in different ways. Comparing the individual impact of different hardware components (e.g., network, storage and memory) over different clusters, we observed 70% improvement in the Hadoop-workload and almost 35% improvement in the Giraph-workload in SwatIII over SuperMikeII by using SSD (thus, increasing the disk I/O rate) and scaling it up in terms of memory (which increases the caching). Then, we provide significant insight on efficient and cost-effective organization of these hardware components. Here, The MicroBrick-based CeresII prototype shows similar performance as SuperMikeII while giving more than 2-times improvement in performance/$ in the entire benchmark test.

Evaluating different distributed-cyber-infrastructure for data and compute intensive scientific application

The superior I/O performance of solid-state storage (eg, solid-state drives) makes it become an attractive replacement for the traditional magnetic storage (eg, harddisk drives) More and more storage systems start to integrate solid-state storage into their architecture To understand the impacts of solid-state storage on the performance of Hadoop applications, we consider a hybrid Hadoop storage system consisting of both HDDs and SSDs, and conduct a series of experiments to evaluate the Hadoop performance under various system configurations We find that the Hadoop performance can be increased almost linearly with the increasing fraction of SSDs in the storage system The improvement is more significant for a larger dataset size In addition, the performance of Hadoop applications running on SSD-dominant storage systems is insensitive to the variations of block size and buffer size, which significantly differs from HDD-dominant storage systems By increasing the fraction of SSDs, there is no need for the Hadoop operators to consider how to carefully tune block size and buffer size to achieve the optimal performance Our findings also indicate that the upgrade of the hadoop storage system can be achieved by increasing the capacity of SSDs linearly according to the scale of the applications

Understanding the Impacts of Solid-State Storage on the Hadoop Performance

To boost the performance of massive data processing, solid-state drives (SSDs) have been used as a kind of cache in the Hadoop system. However, most of existing SSD cache management algorithms are ignorant of the characteristics of upper-level applications. In this paper, we propose a novel SSD cache management algorithm called DSA, which can exploit the application-level data similarity to improve the SSD cache performance in Hadoop. Our algorithm takes both temporal similarity and user similarity in querying behaviors into account. We evaluate the effectiveness of our proposed DSA algorithm in a small-scale Hadoop cluster. Our experimental results show that our algorithm can achieve much better performance than other well-known algorithms (e.g., LRU, FIFO). We also clearly point out the underlying tradeoff between cache performance and SSD deployment cost, and identify a number of key factors that affect SSD cache performance. Our findings can provide useful guidelines on how to effectively integrate SSDs into Hadoop.

Exploiting application-level similarity to improve SSD cache performance in Hadoop

The invention relates to a hybrid storage system based on multidimensional data similarity and a data management method thereof. The hybrid storage system is composed of an upper layer memory unit (3), a middle layer cache unit (2), a lower layer storage unit (1) and a control unit (4). The upper layer memory unit (3) is used for storing system data and data index tables which are frequently accessed in all storage layers. The middle layer cache unit (2) is used for storing part of data which are frequently accessed, and the lower layer storage unit (1) is used for storing all data sets. The control unit (4) is used for being connected with the upper layer memory unit (3), the middle layer cache unit (2) and the lower layer storage unit (1) through I/O ports to constitute the hybrid storage system of a multi-layer storage structure. The invention further discloses the data management method of the hybrid storage system. According to the hybrid storage system based on multidimensional data similarity and the data management method thereof, the accessed speed can be improved, corresponding data can be quickly searched and sought in the accessing process, data access delay is reduced, the data processing efficiency can further be improved, and a request throughput capacity is improved.

Hybrid storage system based on multidimensional data similarity and data management method

The utility model relates to a hybrid storage system based on multidimensional data similarity. The system comprises an upper-layer memory unit (3), a middle-layer caching unit (2), a lower-layer storage unit (1) and a control unit (4). The upper-layer memory unit (3) is used for storing system data and data index tables always being accessed in each storage layer. The middle-layer caching unit (2) is used for storing a part of data which are often accessed. The lower-layer storage unit (1) is used for storing all datasets. The control unit (4) is connected with the upper-layer memory unit (3), the middle-layer caching unit (2) and the lower-layer storage unit (1) through I/O ports to form the hybrid storage system of a multi-layer storage structure. Compared with single HDD storage systems, the hybrid storage system has better data processing performance; meanwhile, by means of the multi-layer storage structure, improvement of performance of the storage system can be achieved at low economic cost; the utilization rate of a cache in the hybrid storage system is improved, and data reading and writing performance is improved through the cache.

Luo Wenhai

Papers

Understanding the Impacts of Solid-State Storage on the Hadoop Performance

Exploiting application-level similarity to improve SSD cache performance in Hadoop

Hybrid storage system based on multidimensional data similarity and data management method

Hybrid storage system based on multidimensional data similarity