Applications of IBM Platform Computing solutions for solving graphic analytics and 3D scalable video cloud transcoder problems

随着大数据时代的兴起，使用高性能计算（HPC）云进行大数据分析的需求急剧增加。IBM Platform Symphony是IBM Platform Computing的一款新软件产品，它提供了一个高性能的运行时引擎，用于在可扩展的共享网格上运行分布式应用程序和大数据分析。它旨在最大限度地发挥计算能力和支持基础设施的潜力，以加速计算成本高昂的应用程序，从数据中提取洞察力，并更快地做出更好的决策。为了在竞争激烈的HPC解决方案市场中吸引客户，IBM对调查和展示其软件在不同应用领域的适用性表示了极大的兴趣。** 我们将研究IBM Platform Computing解决方案（特别是Platform Symphony）的适用性和适应性，以解决三个不同领域的最新问题，这些领域由于需要处理大量数据而需要高计算能力。我们的总体目标有两个方面：（1）测试Platform Symphony在三个不同领域的适用性，以及（2）确定需要对Platform Symphony进行改进以更好地满足IBM客户需求的问题。通过这样做，IBM将吸引更多在分布式计算领域有不同需求的客户。本研究中使用的三个应用问题是：图分析：在非常大的图中进行分布式关键字搜索;数据分析：使用Hadoop中的MapReduce框架从大数据中有效挖掘频繁模式;多媒体通信：用于无线多播的可扩展视频云转码器。

With the rise of big data era, the demand for using high performance computing (HPC) clouds for big data analytics has dramatically increased. IBM Platform Symphony, a new software product by IBM Platform Computing, provides a high performance runtime engine for running distributed applications and big data analytics on a scalable, shared grid. It is designed to maximize the potential of the computing power and the supporting infrastructure to accelerate computationally expensive applications, extract insight from the data, and make better decisions faster. To attract customers in the competitive market for HPC solutions, IBM has expressed a great interest in investigating and demonstrating the applicability of their software in different application domains. **We will investigate the applicability and adaptability of IBM Platform Computing solutions (in particular, Platform Symphony) to solve the state of the art problems in three different domains, which require high computing power due to high volume of data to be processed. Our general objective is two-fold: (1) testing the applicability of Platform Symphony on three diverse domains, and (2) identifying issues on which improvements to Platform Symphony are needed to better serve the needs of IBM customers. By doing so, IBM will attract more customers with different needs in the area of distributed computing. The three application problems used in this investigation are: Graph Analytics: Distributed Keyword Search in Very Large Graphs; Data Analytics: Efficient Mining of Frequent patterns from Big Data using the MapReduce Framework in Hadoop; and Multimedia Communications: Scalable Video Cloud Transcoder for Wireless Multicasts.