Currently, a number of professional data companies, some universities and related oil & gas companies are carrying on research on "big data" application for various fields in petroleum industry, so as to promote the industry overall efficiency. For example, in the upstream fields of the oil and gas industry, with the use of "big data" it can quickly find oil, reduce production costs, enhance drilling safety, and improve yields. "Big Data" is also applicable in the exploration, development, drilling, maintenance and other fields.For example, in the field of exploration, through data analysis and comparison it obtains a more comprehensive and accurate data collection in seismic acquisition process; in the end it achieves more effective exploration results. During the development phase, the big data can help Oil Service Company on evaluating the production process, through data analysis of the geographical environment, enabling more intelligent oil and gas development. In aspect of drilling, at any time through the data monitoring and precaution, it can avoid waste of resources, reduce equipment wear and tear, and reduce the frequency of accidents. In aspect of maintenance of equipment, with the help of prediction by use of big data, certain equipment of high degree of wear get priority and special care during maintenance, so that oilfield operation will also be benefited.

Ultimately, the"big data" in the oil industry reflects three main values. Firstly, the "big data" could help oil exploration and development decisions, to achieve the most scientific decision-making, and to achieve maximum efficiency of production; Secondly, the "big data" can analyze the trend of oil resources and the potential demanding of consumer groups, promote pragmatic innovation and potential market exploitation; And finally, with "Big data" it can monitor and manage production and security of oilfields. Through the network detection, management and field security can be improved.

DataExplosion Incur Changes of HPC Structure and Its Technology Demanding

With the advent of"secondary development" demanding of oil and gas resources, plus speeding up of transition process of oil and gas exploration from land to ocean,since the beginning of 2011, the oil exploration industry entered in the era of big data processing. In order to complete the massive data collection and its in-depth analysis, so as to positioning precisely oil and gas resources, HPC (high performance computing) technology becomes more and more important, and it nearly becomes the backbone of production efficiency ofoil and gas exploration industry.=

Almostevery year, compared to the previous year, there will be an increase exponentially of the data generated in data acquisition segment, and this growth in turn continuously drives the innovation of HPC applications in oil and gas exploration industry. This year, the amount of data explosion again changed the HPC application mode in oil and gas exploration industry.

"Compared to last year, this year the amount of data has increased by three times." says Mr. Nai Neng He, the Chief Engineer of Data Processing Center of Geophysical Research Institute, with the rapid development of seismic acquisition technology and the increase of sampling position amounts, the data volume growth rate has exceeded their forecast in last year. Either conventional processing or high-density processing, today's seismic processing techniques have undergone a lot of changes,from single-faceted collection in the past to nowadays multi-faceted collection with special techniques of the further comprehensive use of its data. The change of data amount over the last year brings great changes of HPC application structure and demanding of related technologies in oil and gas exploration industry.

The first result brought by data explosion is HPC system structure changing.The processing of huge amount of data is very important to parallel optimization and operation efficiency increase of HPC systems. Over the past year, in the Data Processing Center in charged by Mr. Nai, Intel Sandy Bridge V2 structure processors have been widely used in the production system. Collaborative computing platform consisting of CPU enhanced coprocessor had been accepted substantially by oil and gas exploration industry, Mr. Nai said they had tested the collaboration effects between Intel Xeon Phi coprocessor and GPU, compared to Xeon E52670 3.6 times increase of performance as of the test results was very "attractive" to them, and may turns to actual application during this year.

Meanwhile, users’key concern is no longer only to enhance the performance of compute nodes. "HPC is mainly used in mass data processing, but now a single file can be 20TB - 30TB, even 50TB, so requirements will be very high on efficient processing such sale data for the entire computing system,including storage, network, and I / O " For Mr. Nai, it seems that the current HPC challenges in oil exploration industry is not simply rely on the realization of exascale computing capability, for improving the HPC performance it more and more relies on collaboration among the three factors - computing, networking and storage. Besides higher performance computing platforms, during the last two years they mainly invested on the bottleneck of network and storage. Besides improving CPU efficiency and parallel computing performance through improving the algorithm, Mr. Nai regarded that during the past two years the most effective measure of HPC systems modification is the upgrade of memory and storage up to Gigabit data network. Over the past two years, they invested in sum more than RMB 4 million for technological innovation in this regard; the result turns out that "the effect is apparent, with the performance more than doubled."

Moreover, in last year through the use of SSD technology on processing massive amounts data with some key algorithms, Mr. Nai found the production system had obtained efficiency increase by at least six times. Furthermore through the application of Intel DCM data center management software, energy costs also dropped by 10%.

Big Data For Traditional Industries: From Observation To Practice.

Fromtraditional perspective, oil and gas exploration industry is not a typical large data industry, which is determined by the application characteristics. Although it has the characteristic of demanding for massive data processing, its huge amount collected data is hard to get real-time transmission & real-time processing through network, but this cannot prevent users in oil and gas exploration industry from searching their own specific solution among those big data applications, so as to break the barriers of traditional applications.

"Last year, we just got started with Big Data,with eyes on 'how to do'. This year, we eager to know 'how to do it better'." From the perspective of a large data processing, they have found some disciplines, passed through "stumbling" stage on some techniques, no longer as in the past, "one step forward by one step pause." Nowadays it’s piece of cake for them on processing 10TB scale data, in the past they had to spend 9 months on processing 380GB data, now only half that time for dozens of TB.

"Principle jobs of BigData: data collection, data analysis, and utilization after analysis." Said Mr. Gong Yi Min, Senior Manager of Dept. Industry Collaboration & Solution, Intel (China) Co., Ltd, also an expert on industrial structure expert. In his point of view, although the oil and gas exploration industry, unlike the Internet industry which is prone to typical large data applications, actually in many application scenarios in oil and gas exploration industry users can utilize big data related technologies, for example, realizing security through monitoring and big data analysis in oil and gas industry. In the power industry, through prediction with big data analysis, it can solve grid transmission ratio imbalance problem through appropriate corresponding deployment. There are many other such alike application scenarios.

For more information, please visit Zoranoc's official website:www.zoranoc.com