The hottest Microsoft Intelligent Cloud layout hig

  • Detail

According to the data of standard & Poor's capitaliq, IBM's share price fell by 14.2% in 2015, and Oracle investors received a negative return of 18.8%, but Microsoft's share price soared by 19.4%. Among the three new businesses of Microsoft, cloud computing is the top priority. Although Amazon cloud and Microsoft have long been ahead of Gartner's magic quadrant of the global IAAs cloud computing market, in fact, in 2015, Microsoft has begun to focus on the higher-level PAAS market

in the second quarter financial report of Microsoft fiscal year 2016 (as of December 31, 2015) released on January 29, 2016, the revenue of Microsoft smart cloud azure increased by 140% at a fixed exchange rate, among which the revenue of high-end services increased by nearly three times over the same period last year. Since the launch of the first generation azure public cloud service in 2009, Microsoft renamed windowsazure microsoftazure in early 2014, raising the azure cloud business to the strategic level of the whole company. In the middle of 2015, Microsoft again revised the positioning of azure to intelligentcloud. The renaming of smart cloud indicates that Microsoft is getting rid of the competition of IAAs layer and making great strides towards the PAAS layer providing high-end services, that is, the operating system of public cloud

smart connect everything causes data inflation

why do you need an intelligent cloud? Shenyuanqing, general manager of Microsoft China Cloud Computing and enterprise business department, told that IOT will achieve explosive growth from 2016. A world based on M2M (machine to machine communication) needs a new it dimension

according to Gartner, a market research organization, 6.4 billion IOT devices will be used worldwide in 2016. Ericsson predicts that by 2021, there will be about 15billion M2M smart terminals in the world. According to the data released by the International Telecommunication Union (ITU) in May 2015, the number of people in the world will reach 3.2 billion by the end of 2015. In other words, the number of IOT connected devices will be twice that of the number of Internet connected people in 2016, and the gap will expand to five times by 2021

in fact, the development of IOT has greatly benefited from the development of industry 4.0 based on sensors and automatic control in various countries In China, a series of strategies such as made in China 2025 and Internet + are great development opportunities for intelligent manufacturing. The deep integration of mobile Internet and intelligent manufacturing will form a huge IOT

idc once issued a digital universe survey report in 2013, which said that the global digital universe would reach 40zb by 2020, while in 2008, this figure was only 0.5zb. IDC said that the main reason for the expansion of the digital universe is the rapid growth of machine generated data, which will exceed 40% by 2020

in other words, cloud computing, which was born in 2008, was originally an infrastructure designed for the scale of interconnection, but with the continuous development of IOT and the continuous expansion of the digital universe, cloud computing also needs to be transformed and upgraded from the original architecture to a new dimension

global computing resource scheduling

if you have 100000 computing cores (equivalent to 25000 servers, 4 for each), and these packaging materials often become discarded computing cores in the hands of consumers, that is, the most common four core PC servers), what will you do? How about calculating the life insurance cost of everyone on earth?! If it is a single core computer, it will take about 19 years to complete this experiment

at the beginning of 2016, towerwatson, an American life insurance company, and the HPC high-performance computing and bigcompute super large-scale computing group of Microsoft Intelligent Cloud tested this amazing computing project. As a result, Microsoft's cloud computing resources in 14 regions around the world were invoked, and this feat was completed in 12 hours

this experiment not only proves the possibility of global super large-scale computing, but also destroys the Three Outlooks because the huge amount of computing resources extracted are released 12 hours later. Category 1 is the hardness experiment returned to the cloud bases of various countries. During the 12 hours of this experiment, the outside world was also using other azure cloud computing resources at the same time, which was not affected at all

this is a public cloud scenario for extreme Computing: cloud computing provides massive computing resources and powerful scheduling capabilities, allowing global users to use mobile computing resources on a large scale in parallel and concurrently, without hindering other users because a user occupies a large number of computing resources

towerwatson's experiment is, to some extent, an attempt to calculate the digital universe. If the public cloud of IAAs layer stores and manages massive computing resources, the public cloud of PAAS layer is the powerful scheduling system, which is equivalent to the operating system of the digital universe

big data management platform

because of the expectation of the big explosion of the digital universe, Microsoft has evolved from cloud computing in 2008 to today's intelligent cloud, as the management platform of the big data century. Shen Yuanqing said. As the high-performance and massive data testing mentioned at the beginning of this article shows, Microsoft is sparing no effort to lay out the big data operating system: Intelligent Cloud

in an interview with Shen Yuanqing, as the head of Microsoft cloud computing in China, he divided Microsoft's intelligent cloud into six levels: the lowest level is cloud infrastructure, which provides enterprises with public cloud, hybrid cloud and cloud management services; The second is the IOT service layer, which establishes IOT for enterprises through end-to-end suite solutions; The third layer is data + analysis services, which help enterprises conduct intelligent analysis of data on the cloud; The fourth layer is developer + application platform, so that developers can intelligently develop modern applications; The fifth layer is the enterprise mobile layer, which allows enterprises to easily manage SaaS software, mobile applications and user experience; The sixth is the SaaS software layer, which provides enterprises with various SaaS application software

among these six levels, the lowest cloud infrastructure, namely IAAs layer, is the foundation of the entire Microsoft intelligent cloud. Shen Yuanqing said that at present, Microsoft has built 24 data centers around the world, highlighting the layout of Microsoft's global computing scale, which is still increasing. In November 2015, Microsoft just announced the data centers in the UK and Germany. Microsoft currently has more data centers in the world than Google and Amazon combined. Microsoft is also the first cloud computing manufacturer to establish a data center in India

on the basis of IAAs layer, Microsoft intelligent cloud, born in the operating system, has positioned itself on the PAAS platform from the beginning. The so-called PAAS is actually to infinitely expand the operating system of a computer to millions of servers distributed in data centers around the world, and the scale of this virtual operating system can also increase linearly with the number of servers

for the operating system, load balancing and job scheduling are the most basic functions. Corresponding to the example of towerwatson, it can extract 100000 computing cores from Microsoft's global data centers at any time without affecting tens of thousands of other users to use the remaining computing resources at the same time, which is load balancing; When these 100000 computing cores are used, they can be immediately released back to the data centers of various countries, which is job scheduling

such super large-scale computing resource extraction, scheduling, management and recycling on a global scale can obviously not be completed manually, but based on the basic management architecture of Microsoft intelligent cloud, from software definition network to software definition storage to software definition data center, all are realized in the way of machine intelligence and automation

three body world for the future

the PAAS layer of Microsoft Intelligent Cloud includes three levels: IOT, data analysis and application development. Microsoft smart cloud launched more than 600 new functions in the past year, most of which are to enrich and enrich the capabilities of the PAAS layer. Facing the digital universe in the era of IOT in the future, Microsoft also has a larger three body layout: software and hardware integrated intelligent devices, data lake and artificial intelligence

IOT, as the key integrated information infrastructure supporting the intelligent world, can effectively connect machines, people, control systems and information systems, comprehensively, deeply and in real time perceive the data generated by various machines, thus forming intelligent decision-making and providing predictive control, so as to change the operation mode of the whole industry and society

however, so far, intelligent devices belong to the separation of software and hardware, which are produced by software manufacturers and hardware manufacturers respectively, and then combined when used. But in the era of IOT, computing has been pushed into very microscopic physical devices. Due to the very small size of IOT equipment, it requires an integrated solution of software and hardware for power consumption, computing performance, heat dissipation and other requirements

in early 2016, Microsoft announced that ylake, which is deeply integrated with Intel's next-generation chip skylake, is the sixth generation chip that Intel will soon release. At the beginning of the design, engineers from both sides jointly carried out in-depth optimization for IOT and other scenarios, which can achieve 30 times higher image effect and three times longer battery life than current windows7 devices, and support new encryption and security technologies from the bottom of the chip. In August 2015, Microsoft has launched windows10 IOT Core Edition for ultra lightweight devices of IOT devices, which can support 32 or 64 megabytes of ultra-low memory

wuyiheng, director of IOT operating system of Microsoft China Cloud Computing and enterprise division, said that Microsoft intelligent cloud has specially developed different service modules for IOT, so that software developers, system integrators and value-added service providers in different industries can develop customized applications, and finally turn the massive data collected by IOT devices into data oil. These services include azureiotsuite template library, iothub connection center, eventhub event center, streamananalytics flow analysis, machinelearning machine learning, and notificationhub notification center

the iotsuite template library includes two development libraries, remote monitoring and prediction service, for developers to conduct remote two-way communication, remote monitoring, data collection and sorting on the accessed machine data elements, while prediction service learns the historical data through machine learning and artificial intelligence, predicts the possible problems of IOT infrastructure layer according to the results, and then gives early warning. Through these two templates, industry, agriculture, transportation, smart city and other industries can develop IOT solutions for their own scenarios, and incubate a complete industrial IOT ecosystem

in terms of data analysis, Microsoft intelligent cloud can support from enterprise database sqlserver and big data analysis hdinsight to data Lake large-scale big data analysis. What is a data lake? If all raw data are cleaned, processed and processed before real-time data analysis, it can not adapt to the influx of large-scale raw data in the era of IOT, and the cost of data processing is extremely expensive. The data lake stores all the collected original data first, and then calls it in time when data analysis is needed. It is especially suitable for super datasets such as smart cities

in AI, Microsoft intelligent cloud has launched a series of AI computing and analysis solutions suitable for different scenarios, including Cortana business intelligence analysis suite, AI application development platform Oxford plan, data visualization analysis powerbi and AI engine Microsoft Xiaobing. Among them, the Oxford project is to provide developers with a set of APIs and SDK toolkits based on rest and Architecture: including computer vision

Copyright © 2011 JIN SHI