The selection of flash and hard disk under large data storage

With the industrial revolution, the rapid development of science and technology, people choose the storage medium has also undergone a great change, hard disk began to become People's daily work in the necessary products, and in the data storage plays an important role. But in recent years, because of the popularity of social networking sites and intelligent mobile terminals, the amount of data created by people is staggering, according to statistics in 2013, the total amount of data produced in China has exceeded 0.8ZB (equivalent to 800 million TB), under the pressure of massive data, how will the data storage media change? How to upgrade the traditional data center to meet the needs of the business operation facing the massive data? How can emerging internet giants deal with their huge data access problems? These are the problems that need to be solved by traditional means, such as unlimited expansion of data center size, expansion of storage capacity, or the adoption of new scientific and technological means, Select new storage media to meet the need for massive data access?

Data trade-offs under Large data

To solve the above mentioned problems, first of all, we should first understand the characteristics of the current sea volume data. The previous data storage is based on structured data, which has some regularity, can conveniently store the data and can obtain the relevant information at any time through the simple mechanism. But with the popularity of social networks and smart devices, anyone can produce a lot of data that is disorganized and irregular, and is called unstructured data. In particular, the Internet industry, users on its platform to create video, text, pictures are missing a certain information annotation, making the early data center difficult to distinguish between data types, the use of direct storage of data information methods, greatly reducing the availability of data, and caused a certain cost waste.

In the big data age, people began to re-examine the availability and importance of data. How much is the amount of data that is composed of structured data and unstructured data for the cost burden of enterprise IT construction? Can the availability of these data gradually reduce the input-output ratio of the enterprise? If you want to solve these problems, start with the media that stores their data.

Disadvantages of traditional mechanical hard disk storage

We know that since the hard drive has been built by IBM, the blue giant has been shouldering the burden of data storage, in the enterprise IT infrastructure, the hard disk is to build it the overall structure of the "brick." Hard drive manufacturers through the change of technology, manufacturers of various acquisitions, now has formed an oligopoly era, such as the purchase of IBM hard drive business of Hitachi Global Storage by the Western data acquisition, renamed HGST, the Western data is also sitting on the hard drive boss position; Seagate, founded by IBM's former employees, after acquiring the hard drive business of the company and Samsung , but also stabilize the hard disk market second, the hard drive market competition degree can be seen. Because the user's high demand, can create the market prosperity, the share is big, rob people will be more, means will be intense.

So what happened to the high demand for the hard drive in the big data and cloud computing era? Because of the limitation of the traditional mechanical hard disk, the majority of users can only adopt the means of increasing capacity to meet the demand of data access when they face massive data. And the continuous expansion of equipment will inevitably lead to increased energy consumption and space resources waste, which also causes enterprises to fall into a "bottomless capacity, money is not enough to plug" the situation. In addition, because the traditional mechanical hard disk speed, performance and high reliability limitations, so that users access to large-capacity data, increase a certain amount of time costs. Coupled with the diversity of the currently dominant unstructured data, users are no longer content with simple data storage, but need to find data value in thousands of data access processes, which are highly demanding for a single hard drive.

The choice of flash memory

Traditional mechanical hard disk can not meet the new needs of users, the development of technology can not quickly follow the user needs change speed, then everyone began to look for other types of storage media, Flash was raised out. As the name suggests, flash memory is a very good storage medium for access speed. Flash memory has a certain time, but people always reported to it with distrust, in addition to expensive, but also traditional knowledge, small chip storage media is difficult to make the traditional enterprise trust, or a large disc to look at people comfortable and worry ah. So far, many financial institutions are still afraid to use flash memory as the primary storage medium for high-performance computing, but also out of concern for its instability.

But in terms of performance, Flash can definitely address the needs of today's large data age. The speed of access to massive data has exceeded dozens of times times that of traditional mechanical hard drives, and small and refined volumes have saved a lot of space for traditional data centers and saved energy and saved a lot of money for data centers. Of course, the premise of "Saving a lot of money" is that these flash-drives are being sent to you. So, now the application of Flash is still not go out of expensive and unstable "high handsome" image. The two big hard drives have not yet developed solid-state hard drives, which may also be a general direction for market choices.