| Popis: |
Big data poses big challenges to modern enterprises. Traditional data-intensive business applications begin to fall behind the times, because of insufficient capabilities to process large data volumes, ever-changing streaming data and unstructured information. Furthermore, large, complex and all-in-one enterprise applications are no longer popular, whereas lightweight and fragmentary applications become welcome for the sake of the bursts of cloud computing and mobile internet. Below this kind of situation, this paper proposed a big data application platform for enterprises to simply develop and operate personalized data retrieval, data analysis, business intelligence and other data-intensive services. Unlike current related products, it is distinguished for the design of key-value based hybrid data storage and a service-oriented outsourcing architecture. The former is used to resolve the frequently-encountered issue of diversified and massive data storage, and the latter enables an open environment for third-party vendors, which promotes a self-increasing services ecosystem for big data application development. In order to validate the feasibility of our platform, the paper also developed a sentiment analysis application for Kingdee popular products based on the services provided by the platform. The result of our experiment proves that the platform shows enough potential to effectively facilitate data-intensive application development. |
