HuazhangUniversity of Science and Technology is one of the earliest universities to conduct research on big data. In order to provide innovation environmentwith the transformation of research results to practical application, speed up the transformation of the large data application technology to research direction and promote the theoretical research of the big data, Huazhang University of Science and Technology sets out to create the big data collaborative innovation platform.
Mass data storage: The research contents of embedded and pervasive computing laboratory of Huazhang University of Science and Technology involves the Internet of things, Mobile application and Robots, and there are a lot of sensors and Data Acquisition Equipment (DAE) produced. Andin the process of operation, the equipment generates a lot of datum that includes pictures, video and audio data besides structured data. Adopting traditional relational data is very difficult to realize quick warehousing of mass data and it also cannot give strong support for semi - structured or unstructured data.
Mass data analysis: the core of the Internet of things and artificial intelligence research is analysis and mining of all kinds of data. High performance clusters provided by universities are aimed at compute-intensive applications and they cannot do well in big data analysis characterized by IO-intensive, so they are difficult to meet all kinds of various research requirements.
The infrastructure deployment and maintenance: different from IT department of traditional companies, team of embedded and pervasive computing laboratory has strong professional ability and academic levels. But the team is difficult to solve problems quickly because the team lacksrelevant experience in IT infrastructure maintenance, design of big data analysis platform and tuning and management, which can take up most valuable time and energy of scientific research.
We adopt InspurSmartDate Appliance as the infrastructure layer of the collaborative innovation platform. Datum produced by sensors, webcams and mobile devices are collected in the InspurSmartDate Appliance by the data acquisition engine and they are recorded quickly in the specialized file system of the InspurSmartDate Appliance.
Distributed computing engine processes the datum according to needs of different applications and eventually various kinds of data analysis models and technologies process the datum to support the scientific research needs in the fields of statistical analysis, artificial intelligence and datavisualization.
Full data storage and processing:InspurSmartDate Appliance adopts full-distributed architecture and has extremely highaggregate storage broadband, which can guarantee that mass data is recorded quickly. Inspur’s enterprise-class hadoop distribution packed with it can support various kinds of storage, processing and analysis.
Rapid business response: InspurSmartDate Appliance can make a whole tuning of software and hardware according to applications of different fields, perform well in mass data query, achieve second-level return of ten billion data queries and help clients to improve the speed of business response.
High-expansibility: InspurSmartDate Appliance adopts the distributed architecture. With the increasing of business, system can increase elastically and achieve linear expansion following the expansibility of the systemand help clients to realize the real high-expansibility.
Speeding up business delivery: the overall delivery of software and hardware of InspurSmartDate Appliance solves compatible adapter problems of different brands, software and hardware about which traditional business deployment needs to think, and plug-and –play can reduce greatly the time of business deployment so that achieve the real plug-and-play. InspurSmartDate Appliance with the integration of software and hardware can decrease the time of system deployment and accelerate the speed of the business delivery.