Ten-year outlook: Big data surged 50-fold and server tenfold

For the first time in 2011, big data of 1.8ZB (1.8 trillion GB) capacity was born for the first time, equivalent to one American writing 3 tweets a minute, and continuously writing 26,976 million years. According to IDC's forecast, global big data capacity will surge 50 times over the next decade, while the number of servers managing data warehouses will increase by more than 10 times to meet demand.

Undoubtedly, big data is about to challenge enterprise storage architectures and data center infrastructures, which will also trigger a ripple effect among applications such as cloud computing, data warehousing, data mining and business intelligence.

We know that there are many problems in the construction of the application system in the Internet. For example, the customer base is uncertain, the system size is uncertain, the system investment is not fixed, the business application has a clear parallel segmentation feature, the construction of data warehouse system, the data warehouse The scale can be estimated, the data warehouse system investment and business analysis of the value and return, business intelligence applications are part of the overall application, Saas mode to build a data warehouse system.

It is possible to build a next-generation data warehouse using cloud computing in big data technology. From the perspective of system requirements, the architecture of big data poses new challenges to the system, which can also be regarded as the "godsend opportunity" for cloud computing:

First, the higher integration: This means that a standard chassis to complete a specific task, Asus will soon introduce a high-density rack-mounted server RS720, 2U height can be used to support four dual computing nodes, to achieve stand-alone 8 Intel 5600 series processors and up to a total of 768G memory resources.

Second, the configuration is more reasonable and faster: storage, controller, I / O channel, memory, CPU, network balanced design, optimal design for data warehouse access, more than an order of magnitude more than traditional similar platforms, this classic The case is Teradata, a data warehousing leader, with an enterprise-class data warehouse with a dual Xeon six-core processor. The 5650 easily handles more complex, larger workloads, sustained workloads and batch loads, operational inquiries, Simple reporting and complex analysis, all functions run on the same platform. Compared to the previous generation, the dynamic enterprise data warehouse 5650 delivers up to 43% better performance with a reduced floor space and reduced energy consumption and space requirements.

Third, the overall lower energy consumption: the same computing tasks, the lowest energy consumption.

Fourth, the system is more stable and reliable: to eliminate all kinds of single point of failure, a unified component, device quality and standards.

Fifth, the management and maintenance costs are low: the data collection management of all conventional integration.

Six, plan and foreseeable system expansion, upgrade road map.