Cloud computing and big data are two sides of a coin

In the age of mobile interconnection, tens of billions of machines, businesses, and individuals will acquire and produce new data anytime and anywhere

Even with the support of Moore's law, which increases the performance of chips by 1 time-fold every 18 months, the pace of hardware performance evolution is already lagging behind data growth and the gap is growing.

Within 1 minutes, Sina Weibo sent tens of thousands of microblogging, Apple App store downloads tens of thousands of times, Taobao sold tens of thousands of items, Baidu produced million search queries ... All of these behaviors are presented by massive amounts of data.

December 12 last year, the promotion period, Taobao launched "Time Machine"-a few years Taobao buyers according to the purchase of merchandise records, browse clicks, receiving address and other data editing "personal network to buy a blog," so as to record and outline the memories of life. Behind, is based on 470 million Taobao registered users online purchase data analysis and processing, which is the typical application of large data.

With the development of traditional internet to mobile interconnection, globally, in addition to common computing terminals such as personal computers, tablets, smartphones and game consoles, broader, interconnected smart devices such as smart cars, smart TVs, industrial equipment and handheld devices are connected to the network. Based on the platform and application of social networks, tens of billions of machines, businesses and individuals can acquire and produce new data anytime and anywhere.

Internet search engine is one of the most typical applications of large data. Baidu daily Processing data volume reached dozens of PB, and presented a trend of rapid growth. If a CD-ROM capacity is 1GB, this is equivalent to tens of millions of discs on a base. Microsoft Bing, a search engine in China, needs to respond to 10 billion-magnitude search requests a week. By working with Facebook, more than 1 billion social network search requests are processed every day via Bing.

China Mobile's Internet traffic has increased 10 times-fold in just 18 months. Hequan, an academician of the Chinese Academy of Engineering, said that as social networks matured and mobile bandwidth increased rapidly, more sensors, mobile terminals, more data and more growth than at any time in history, data traffic on the Internet was growing rapidly. Hequan that China's mobile internet has moved into the "Big data" era, driven by technologies such as cloud computing and things networking.

According to IDC, a market-research firm, the total amount of global information will grow by one times every two years, with a total of 1.8ZB (1ZB about million PB) produced globally in 2011, compared with 1ZB in 2010, which is equivalent to the sum of global historical data.

After cloud computing, big data is one of the most popular concepts in the field of information technology.

Large data has four characteristics, the most important is to gain insight and value

In the IT industry, a large data industry has been defined as: "Information services based on data storage, value extraction, intelligent processing and distribution on the basis of a wide range of data sources, such as the Internet and the IoT," or, as the IT giant outlines big data strategies: " Dedicated to enabling all users to gain insights from virtually any data that can be translated into business execution, including insights that were previously hidden in unstructured data.

"In short, a large, dynamic, sustainable data, through the use of new systems, new tools, new models of mining, so as to gain insight and new value of things." "Zhang, Microsoft's global senior vice president and chairman of Microsoft Asia Pacific Research and Development Group, said in an interview.

Although there are many interpretations, the industry generally considers large data to have four "V" characters that begin with: Volume (volume), produced (kind), velocity (velocity) and the most important value. Volume refers to large data volumes and data integrity. Zhang said that the data it industry refers to, the birth of more than 60 years. And until the popularity of personal computers, due to the storage, calculation and analysis tools of technology and cost constraints, many natural and human society is worth recording signals, does not form data. A few decades ago, meteorological, geological, petroleum geophysical, publishing, media and film industry is a large, continuous output signal of the industry, but at that time more than 90% of the use of storage analog signals, it is difficult to use computing equipment and software for direct analysis. Governments and enterprises with a large amount of funds and talents can only extract, transform and load a few of the most critical signals into the database.

Zhang that, although the industry on how to achieve the magnitude of the large data is not conclusive, but in many industries in the application scenario, the size of the dataset itself is not the most important, whether integrity is the most important.

Produced means discovering an intrinsic connection between a vast and varied range of data. In the internet age, a variety of equipment through the network to become a whole. Entering the Web2.0 era characterized by interaction, personal computer users can not only obtain information through the network, but also become the manufacturer and disseminator of information. At this stage, not only is the volume of data starting to explode, but the range of data is beginning to grow.

"This necessarily prompts us to analyze, process and integrate massive amounts of data, to find the ' relevancy ' of data that would otherwise seem irrelevant, and to turn seemingly useless data into useful information to support our judgments." "Zhang said.

Velocity can be understood to meet real-time demand faster. The need for real-time data is becoming clearer. For the average person, driving to dinner will first use the map in the mobile terminal to check the location of the restaurant, anticipate traffic congestion on the road, get information about the parking lot and even comment on the restaurant by other users. When you eat, you can use your cell phone to take photos of your food, edit a short comment, post it on Weibo or micro-mail, and use lbs (location-based services) to find people who eat at the same restaurant and see if there are any friends nearby ...

Now, with all kinds of wired and wireless networks, there is a ubiquitous connection between people and people, people and machines, machines and machines, and these connections inevitably bring data exchange, Zhang said. The key to data exchange is to reduce latency to present to users in a way that is near real-time-meaning less than 250 milliseconds.

But more important than the previous 3 ' V ' is value, which is the ultimate meaning of large data--gain insight and value. "The rise of big data is driven by the rapid development of technologies such as artificial intelligence, machine learning, and data mining," Zhang said, presenting a process that transforms signals into data, analyzes data into information, refines information into knowledge, and makes decisions and actions with knowledge.

Baidu related experts believe that, in terms of the value of large data, just like the sand gold, the larger the size of large data, the real value of the relatively few data.

"So really good big data systems, the more important is not the better, in fact, the less the better." "Zhang said, the first data to be more, the best or less, the ZB, Pb eventually become a bit, that is, the final decision." That's the key.

Cloud computing and big data are two sides of a coin, and big data is triggering profound technological and commercial change worldwide.

Like the advent of cloud computing, big data is not a sudden new concept.

"Cloud computing and big data are two sides of a coin, cloud computing is the IT Foundation for big data, and big data is a killer app for cloud computing." "Zhang said. Cloud computing is the driving force for big data growth, and on the other hand, because the data is more and more complex and more real-time, it needs cloud computing to deal with, so the two are complementary.

30 years ago, the cost of storing 1TB, which was about 1000GB, was about $1.6 billion trillion, and now it takes less than 100 dollars to store on the cloud, but stored data, if not mined and analyzed by cloud computing, is only zombie data, not much value.

At present, cloud computing has become popular and becoming the mainstream technology in IT industry, which is a kind of infrastructure and business model which is born in the background of more and more computing, more and more data, more and more dynamic and more and more real-time demand. Individual users upload documents, photos, videos, game archive records to the "cloud" permanent preservation, enterprise customers according to their own needs, can build their own "private Cloud", or hosted, or rented "public cloud" on the IT resources and services, these are not new. It can be said that the cloud is a tree full of large data of the apple tree.

The emergence of large data is triggering profound technological and commercial changes worldwide. In technology, large data makes the conventional way of extracting information from data changed. "In the field of technology, more rely on the model of the method, now we can borrow a large scale of data, using a statistical method, is expected to enable speech recognition, machine translation technology areas in the large data age to make new progress." "Zhang said.

Machine learning, which plays an important role in search engines and online advertising, is considered an area where large data play a real value. Statistical analysis of human behavior, habits and other methods in a large number of data, the computer can better learn to simulate human intelligence. With the increasing popularity of natural user interfaces, including voice, vision, gestures, and multi-touch, computing systems are being able to perceive, understand and understand human users in a way that is similar to humans. The increasing perceptual ability of this computing system, combined with the advances in large data and machine learning, has enabled the current computing system to begin to understand the intentions and contexts of human users. "This allows the computer to really help us, or even to work on our behalf."

In the business model, Zhang that large data means exciting opportunities for business and service innovation for business competition participants. Retail chain Enterprises, electric business giants have been in large data mining and marketing innovation has a lot of success stories, they are very sensitive to business acumen, the courage to invest in the future of the company, and thus obtain a generous return.

IT industry chain division, the dominant power also because of large data has a huge impact. In the past, mobile operators and Internet service operators have a large number of user behavior habits of various data, in the IT industry chain has a pivotal position. In the big Data age, mobile operators could be piped thoroughly if they could not dig out the value of the data. There has been a consensus among operators and third party developers who know more about the needs of users.

(Responsible editor: Schpeppen)