Hadoop alone is not enough for big data processing.

Since Oracle, IBM, and Microsoft announced their support for hadoop in 2011, hadoop has basically become synonymous with big data in the enterprise-level market. Today, this situation may change.

Nosql is becoming increasingly important

Thanks to the high profile of hadoop, few people have noticed that the three major relational database vendors also announced their support for non-relational database nosql in the same year that they announced their support for hadoop.

As an open-source software, nosql (not only SQL, not just SQL) was born and developed to meet the "Three highs" demand of databases in Web 2.0, especially social networks, that is, high-concurrency reads and writes to the database, efficient storage and access to massive data, and high scalability to the database. These are the shortcomings of relational data.

However, to apply nosql on the Internet to the enterprise-level market, users cannot be completely put down. After all, the Internet rarely runs key business applications, and which Internet giant has not interrupted services, but downtime is more embarrassing for Internet companies. However, for enterprise-level users who put their personal lives on it, their key business applications cannot be lost.

"Oracle nosql databases are divided into free and paid versions, providing commercial support at the same time. As a scalable and highly available key-value database, Oracle nosql database integrates with Oracle event processing to provide scalable key/value storage, allowing you to efficiently capture real-time big data ." Yu sicheng, vice president of Oracle and general manager of technology in Greater China, said,

According to Yu sicheng, based on the development in recent years, the paid commercial version is more recognized by customers at the enterprise-level application level. For example, visa recently deployed nosql to detect fraud. "When many enterprises develop to the hadoop level or bring some value to big data technology, the industry needs a vendor to fully support all the above architectures, this also proves Oracle's leadership in the big data layer. With the coexistence and integration of hadoop, nosql, and SQL, big data technology will have a higher maturity ."

Overall Big Data Solutions

"Structured data and unstructured data are mutually existent and inclusive, and the integration of the two processes constitutes a complete Big Data Processing ." Yu sicheng believes that "after Sun's acquisition, Oracle has the ability to integrate hardware and software, and then form an integrated system to provide customers with solutions with complete big data processing capabilities ."

According to reports, nosql is integrated with Oracle event processing (OEP) to efficiently capture real-time big data and then process and query it through mapreduce. The real-time filtering, association, and event processing applications built by OEP can help eliminate, integrate, associate, and filter data, avoid data warehouse clutter, and thus have the ability to analyze a large amount of data streams in real time.

　　After the emergence of hadoop, Oracle built a big data processing platform through hadoop Enterprise Edition and Oracle nosql databases.

Later, Oracle formed a database cloud platform for Oracle exadata to process structured data by integrating database engines (DBE), storage, servers, and optical fibers. The integration between hadoop and Oracle database and exadata is optimized through the Oracle big data connector. The hadoop system is seamlessly connected to the Oracle database system for the stitching of unstructured and structured data, to form a complete big data solution.

In terms of data analysis, Oracle exalytics Business Intelligence ECs is developed using industry-standard hardware, business intelligence software, and memory database technology, it is the world's first integrated system designed to provide high-performance analysis, modeling, discovery, and planning. It loads and reads data from Oracle exadata through the ultra-high bandwidth InfiniBand bus.

In the decision support stage, information is loaded from different source systems and stored in a faceted data model that dynamically supports changed data. As a powerful tool for enterprise information exploration, Oracle endeca information discover is used for advanced and intuitive exploration and association analysis of these multidimensional and variable data.

"With its comprehensive, hardware and software integration, high performance, and high cost-effectiveness, Oracle's big data solution has been supported by many enterprise users ." Yu sicheng said.

Hadoop alone is not enough for big data processing.