Brief Description Data Cloud: The birth and will coexist with the database in harmony

On the afternoon of August 8, 2012, the VMware & EMC Big Data Cloud Summit officially kicked off. At the forum, Fan Chengong, VMware's senior vice president, shared with participants about the changes in data management technologies. The data, cloud, and Changes in cloud users have resulted in three major trends in cloud development. These three major trends have created a data cloud that will coexist with traditional databases for a long time. Virtualization technology and open source software will create a bridge that connects both parties Coexistence of the bridge.

In the background of the current big data boom, VMware and EMC jointly organized the forum to grasp the development trend of cloud computing and big data technologies, combined with the advantages of big data technologies such as Hadoop and cloud computing, introduced its Complete big data cloud solution, and share many customer success stories.

Fan Chengong from the technical point of view, analysis of the five major trends in data management technology changes, data acquisition, analysis of changes in the way, and VMware in cloud computing and big data solutions. This article will introduce you to these five major trends, as well as the five major trends may bring changes in the market structure.

Three trends gave birth to data cloud

The database is now one of the most advanced data management models that allow for good classification of data, relational databases that organize data for quick retrieval, and transactional processing. Because of the generation of relational databases, many applications are now very smooth development.

In the past 10 years, data warehouse has developed very fast, you can a large number of deep-level technology to explore, in the database technology to provide customers with higher value.

These mean that it is hard for users to replace the database without changing the application, and it is not easy to migrate data from one database to another.

However, Fan Chenggong believes that in the past five years, there are some changes in this situation. He said that there are five relatively large trends that will make the situation of a dominance database unified data management will have a more fundamental subversion.

The first is the data itself changes.

Most of the past data is artificially generated, its data is a record-based, more easily converted to a relational database. The treatment of it is often not real-time, you can wait for the data generated, and to use it, it has often been a while. In this case, the relational database is a very good digitization. To take a very simple example, I went skiing at the beginning of the year and went for a sleigh ride. I went to a husband and wife shop, a small sled shop. They did not have a computer and did not have a database. They recorded everything on paper and a pen. . I see that they have a box of cards, each transaction is a card. There is another box is all his customer information, the middle of the customer's information can check each transaction information. When I look at it again, this is a relational database made of paper. If the business is good and the scale is large, it can not be done with paper and pen. Instead, it must be made into a database on a computer. This database has several features, we all know CRUD, need to be able to ensure the data generated, there are data to read and write and change, but also to ensure that data can be deleted, which is the so-called record-type data. The management of such data, the database is a very good, very perfect technology. And now the source of the data is more, and much of our data is no longer generated by humans, but machine-generated. With the development of the Internet of Things, a large variety of detectors, various types of RFID, various mobile handsets, various devices, and many computers, the server automatically generates large amounts of data It is often produced in the form of a stream. Even man-made, including the social networks we just mentioned, weibo, the form of the data is somewhat different from the past.

We see that new data often seldom changes what has been produced in the past. These data are often generated once and never changed again. A server log will no longer change yesterday's log, I put a microblogging yesterday, it will not be changed, often once the data will not change. And many of these data will not be deleted, even if the user to delete it, often in the underlying infrastructure which is not deleted. Under the emerging data we noticed that the CRAP data model is generated, repeated, copied, can be added, but also must be integrated. It is such a large-scale flow-mode data generation, but at the same time it should be a good induction and integration. For such data, we are familiar with the relational database is no longer the best technology to meet his needs. In dealing with such large CRAP data, we need new data management technologies and products to help customers solve this problem. That's why popular technology like Hadoop is now included because past data is no longer sufficient for new big data CRAP data.

The second is the cloud side effects.

The cloud is that your application is not just behind your firewall. With the emergence of software as a service, we will live in a cloud of life. For many businesses, many of our applications are in private clouds and in their own data centers. At the same time, however, more and more of our applications are made available to the public cloud, including client management, including personnel management, and even financial management later, all through the public cloud. And this has a side effect, is that the data is often used together. When your application is outside your firewall, its data is outside the firewall. I am available to you as a software as a service provider, and the data is available to me. As an enterprise, for the first time in the face of this situation, the company's data is not completely controlled by me, I can not put all the data in the Oracle database. Even if I as a CIO have such a wish, can not reach this reality. Because in the end this application which database to use, it is not by my IT department has the final say.

In such a multi-site and multi-sourced data era, how to uniformly analyze and process these different data types and different data materials and get intelligence from it is a new generation of challenges. In the past to be a new application, as long as the connection to the existing database on the line. And now there must be a global unified cloud data system in order to be able to develop new applications that will allow it to extract data from your private cloud as well as data from the public cloud. So this is another trend brought by the cloud, making the data management model will have a more fundamental change.