EMC: Big data = big opportunities

The big data age has come

The difference between big data and traditional data is the expansion of data volume. According to the global data volume tracked by EMC and IDC, the digital world will grow 44 times times in the last 10 years, from 0.9 ZB to 35.2 ZB. Second, the type and complexity of the data will be greatly increased, no longer only to deal with the internal structure of the data, more http://www.aliyun.com/zixun/aggregation/13739.html "> unstructured data, as well as external data.

Unstructured data will account for up to 90%, and industries and applications such as mobile sensors, social networks, electronic payments, video surveillance, video rendering, smart grid, geophysical prospecting, medical imaging, genetic sequencing, and so on, will produce a large amount of unstructured data. These data are highly demanding for real-time. For example, a two-day microblog data may be less valuable or worthless to people. Such a complex, broad data, as a business should be how to deal with it? If the value of the data can not be extracted, then the data for enterprises and individuals is cost, because it requires storage, management. Only by finding additional value, more expensive value than the storage itself, can large data create value for the enterprise.

Big Data trip

Like a few years ago, when companies embarked on a cloud computing journey, EMC gave a few steps to the cloud-computing journey. Similarly, large data is made up of a few steps: first, the transformation of existing IT architectures, including the large data base 820.html "> Storage architecture and Data analysis architecture, to meet all the data needs of faster response and flexible scalability."

Large data environments and traditional schemas are very different for storage and data management. The weaknesses of the traditional pattern of it architecture and data analysis have emerged. First, the scale-up mode, the process of vertically expanding, must introduce new storage systems when the original storage capacity exceeds, and all new provisioning is manual. In the case of large data, both the cost, the data needs of the response, vertical expansion and manual mode are not appropriate. Requires Scale-out mode, automatic provisioning. Second, in the traditional mode, it is easy to form storage islands, data islands, many capacity is either not released in the island, or need more administrators, so that the management structure becomes more complex. In the case of large data, a large storage pool is required, and the storage space of the pool is allocated according to different data.

EMC Isilon

EMC Isilon is a technology for large data, with unprecedented scalability, unprecedented capacity and exceptional ease of operation to manage 15PB. Also can maintain good file system IO/S performance, IO operation may reach millions. More important is the ease of operation, each time the need for expansion, the enterprise according to the data volume growth process, only need to add new nodes.

New Large data analysis platform

In the field of data analysis, large data and traditional data age are very different. Traditional data analysis is limited to structured data and analyzes TB-level stale data. The entire analysis system is limited to the vertical extension of the architecture. As the volume of data increases, different ratings and upgrades are required.

In the case of large data analysis, the first must be a replacement, which can handle structured and unstructured data well. Hadoop is a good technique for dealing with unstructured data, and a good large data platform must be able to handle both structured and unstructured data. Second, to have a good throughput capacity, to handle the PB level of data, to ensure that the data analyzed, master the results are now the most real-time results, to help enterprises make the right choice. The Scale-out architecture is the only option.

(Responsible editor: The good of the Legacy)