Patricia Florissi: Large data needs are real

Now, the topic of big data is even more than cloud computing. Patricia Florissi, Vice President of EMC Inc. and global chief technology officer for sales, said: "At present, we do not have a profound understanding of the benefits that large data can bring to people and how far they can affect people's lives and work." Now, I go to the Amazon to buy books, will not only buy a book, but according to the site's recommendations, buy some other things I am interested in. This is the new changes that the big data age brings to people's lives. ”

Large data needs are real

The big data is a change, it not only affects people's life, work, and more importantly, it affects the way people think about problems. Many people think that the main role of large data is to help manufacturers more accurate understanding of consumer behavior, such as the purchase of a brand of mobile phone users will usually buy another brand of clothing. In fact, the function of large data is much more than these, large data will affect people's decision and behavior pattern to a great extent.

Patricia Florissi told reporters: "Through communication with customers, we find that many customers do not understand the value of large data in the end." But regardless of whether the customer is now understanding the content of large data, the vast majority of customers face large data will not stand idly by. Many companies have at least one department or one person doing things that are related to big data. ”

When cloud computing was emerging, many people were keen to discuss whether cloud computing was a transformative innovation, a "new bottle of New wine" or "a new bottle of old wine". Is there a similar problem with big data? Cloud computing is changing it's consumption patterns, and big data is changing the way we work, live and think. Patricia Florissi that large data is not just a large amount of data, but represents three new trends: first, massive data changes the way people look at things and look at data; second, because of cloud computing, people have the ability to deploy larger amounts of storage and have a stronger ability to handle large volumes of data; Third, people already have a certain level of knowledge and technology, can carry out large data analysis.

Patricia Florissi The example of a printing press. The printing press was invented in the 15th century, but a large number of commercial applications of the printing press took place hundreds of years later. When the printing press first appeared, although the ability to quickly print out a lot of books, but then there are not many people reading and writing. With the progress of human civilization, people have mastered more cultural knowledge, the printing press is really useful. Large data processing and analysis will also experience such a development process. When large data appears, if people do not have enough storage space and storage capacity, then you can only delete a large amount of data, if people have enough storage capacity, but do not have the ability to analyze data, then large data is not worth it, like the face of a large number of gold ore, but can not extract gold from it; If you have storage capabilities and data analysis capabilities, but people do not have the ability to interpret data, then can not be mined the value of data. "Now that we have storage capabilities, data analysis capabilities, and data interpretation capabilities, large data applications are not illusory but real." Patricia Florissi said.

Real-time processing power is more important

Many people first think about the processing of unstructured data when it comes to big data. IDC's statistics show that unstructured data already accounts for the 80%~90% of total data. Therefore, the processing of diverse data has become the focus of many users. But some storage vendors believe that, over time, users will no longer care whether the data is structured or unstructured, because real-time data processing is the key to large data processing.

Patricia Florissi that the user's need for real-time data processing is becoming increasingly urgent. The focus on real-time data analysis has gone beyond the focus on the accuracy of the data itself, which is the result of a surge in data volume. "If users have only a small number of data or data samples, then the accuracy of the data is very important to the user, if the user is faced with massive data, then the accuracy of the data is no longer so important, because a lot of data can make up for the lack of data accuracy." "The value of the data is Patricia," says Florissi. For example, I got a coupon for the mall today and I can use this coupon to buy a dress tomorrow, and no one will pay any heed to it for weeks or months. ”

EMC now has the technology and the ability to find some of the structural features of unstructured data that can be used to process and manage unstructured data for some analytical methods and techniques for structured data.

Storage can be distorted

At the upcoming EMC World Conference, EMC will release its "Software definition storage" (SDS) new product. However, before the new product was officially released, Patricia Florissi refused to disclose the technical details of the product to reporters.

Patricia Florissi believes that SDS will subvert the existing storage market, software definition Network (SDN) will subvert the existing network world, software Definition Data Center (SDDC) will subvert the existing data center market. These changes will come together as a powerful force to subvert the entire IT market.

"Software definition" means intelligent steering software from hardware. For example, the use of mobile phones, users will be in accordance with their preferences in the mobile phone installation of different applications, so everyone's mobile phone is different. The software definition gives the handset a new ability to define hardware configuration flexibly through software. Another example, if the previous user purchased more than one network device, they had to use manpower to configure each network device, not only time-consuming and laborious, and no matter how the device configuration changes, the basic functions of network equipment will not be changed. If the user uses the SDN, only very few hardware, can design the network to be oneself needs appearance, lets the network equipment have the user to want the function. Hardware is like a piece of clay, it can be modeled according to the user's needs, through the software into various forms.

"From a storage point of view, the previous storage device division of labor is clear, file storage can only store files, block storage can only store block data." After the concept of SDS emerges, users can store and manage all files, block data and object data uniformly on a unified storage platform. Patricia Florissi says, "Storage virtualization is just a subset of SDS." Through virtualization technology, users can build virtual storage pool of file or block data, and SDS constructs a comprehensive data pool, which can assign different functions of users to different software levels. ”

Storage management includes both control management and data management: control management determines which information block the data is stored in, and the task of data management is to migrate the data to a suitable place as quickly as possible, two different functions. In software-defined storage, the part of the hardware that performs control management functions becomes less important and is simply a cheap storage medium. Because specialized data needs to be stored at a very fast rate, the requirement of hardware specialization is more and more high at the level of data management.

Pivotal is a new starting point.

At the beginning of 2013, EMC and VMware jointly established a company--pivotal focused on large data and cloud computing business. The company's establishment will help EMC to further promote the implementation and development of its large data overall strategy. Patricia Florissi says: "EMC will focus on storage and data management in the future, cloud computing, large data and trusted computing will become the company's three strategic core." To meet the needs of cloud computing and large data applications, storage must be further enhanced in terms of flexibility, economics, and availability. ”

Specific to large data processing, EMC is more concerned with how to extend storage, including scaling and vertical scaling, while also considering the rapid and secure movement of information between different storage tiers to optimize information movement. Users should also focus on how to back up and archive large data. In addition to the concept of large data, the industry is now hotly discussing a new concept-fast data, whose goal is to deal with massive amounts of data at a very fast rate. In order to improve the efficiency of data processing, the processing and analysis should be close to the data, that is, data processing in the data generation.

Patricia Florissi concludes: "In large data, EMC and pivotal Division of labor clear, pivotal major data analysis, and EMC's core business is large data management." ”