Source | Science and Technology Orientation Theory (xiangling08 15)
Under the wave of new infrastructure, data and computing are becoming more and more important as the supporting force at the bottom.
Recently, sponsored by China Big Data and Intelligent Computing Industry Alliance, the second China Supercomputing Conference ChinaSC with the theme of "New Computing Power, New Infrastructure and New Economy" was held in Beijing. More than 600 people, including academicians, well-known scholars and industrial café s at home and abroad, participated to discuss the progress of cutting-edge technologies such as supercomputing, new infrastructure, cloud computing, big data, artificial intelligence and blockchain.
The official criteria for awarding this award are: "Being able to organically integrate various current technologies, meet various comprehensive software, hardware and system solutions in different application scenarios, and integrate many practical indicators such as scientificity, advancement, stability and economy, which is the key link for technology to be transformed into practical applications."
Obviously, this authoritative award is most concerned with the ability of big data solutions to promote the transformation of technology into practical applications, which is also the core demand for big data in the current market environment. The author tries to disassemble the layout of Kunpeng's big data solution from the macro level to the operational level, hoping to give reference to relevant practitioners in this field.
Technical competition continues, but the demand for big data has turned to applications.
The value of data is becoming more and more obvious, and the technology to better release the value of data is also developing. However, as more and more government and enterprise organizations begin to use big data capabilities to help improve their physical business, their needs are more inclined to whether technology can be better applied, and big data solutions are born for this.
Taking Kun Peng as an example, in the process of promoting technology landing, its big data solution presents three characteristics that meet the needs of the times, making it lead the new trend, widely welcomed by customers and recognized by the authority of ChinaSC.
1, ultra-high performance is still the most powerful support for application landing.
The premise and support of big data solution to promote technology to realize various scenarios is that the performance of the underlying software and hardware itself should be strong enough. Otherwise, without performance support, the perfect and in-depth solution can only be a castle in the air.
Only when the performance is strong enough can we meet the various data value needs of customers as much as possible in the application landing stage.
Thanks to the in-depth development of the underlying software and hardware capabilities, Kunpeng's big data solution has ultra-high performance and is fully prepared and supported for the application.
For example, in terms of hardware, we use the self-developed high-performance Kunpeng 920 processor. In terms of software, we have original IO intelligent prefetching and Spark machine learning that can double the performance in big data scenarios. Graphic enhancement algorithm.
Take the cooperation between Kunpeng and Zhejiang Mobile as an example. In 20 19, Zhejiang Mobile has successively completed the testing of the IT cloud Kunpeng server, the verification and commercialization of the front desk system, CRM, billing, big data, CDN and other systems. Among them, the CRM of Zhejiang Mobile &; With the support of Kun Peng's big data solution, the overall BOSS system has been greatly improved. The system has been running stably for a year under the condition that the scale carries the network operation support business.
At present, Zhejiang Mobile has built the world's first ICT full scene model point in the field of operators around network cloud, IT cloud and mobile cloud.
2. Full-stack solution can promote the comprehensive landing of technology.
The solution itself is not a specific technology, its value lies in the effective integration of various technologies, and it is a unified output model for government and enterprise customers. With the increasing demand of government and enterprise customers, it is necessary to establish a stack system as complete as possible, meet the solutions of various scenarios and technical requirements as much as possible, and integrate various technologies organically, systematically and comprehensively.
This is the architecture of Huawei Kunpeng's big data solution. Based on Kunpeng's processor, an end-to-end full-stack system connecting hardware, operating system, middleware and big data software is constructed, and the full-stack performance is optimized accordingly to promote the integration of various technologies into high-performance solutions:
It can be seen that this full-stack system, on the one hand, can easily meet the needs of science, advancement, stability and economy at the same time through organic integration (for example, acceleration features and big data components can help the solution to be more efficient and lower in cost); On the other hand, as a comprehensive, complete and integrated information solution, it is easier to adapt to the application needs of different industries such as government, finance, telecommunications, Internet and large enterprises.
From technology to application, "full stack" has become an important intermediate conversion link, which is not only "lossless" but also "gain".
3. Meet the personalized needs of government and enterprises, and make technology landing more practical.
When it finally landed as a single customer, big data solutions need to truly meet the actual needs of this customer, that is, a "one-stop shop" from technology to application. After all, no matter how powerful the performance is, no matter how perfect and flexible the full-stack system is, it still needs to conform to the actual business and generate actual value.
It is necessary to have the ability and meet the needs. Kunpeng's big data solution is to do this.
20 19 Jiangsu province built the first provincial, autonomous region, city and county government big data based on Kunpeng architecture. In the future, more and more government affairs systems can be carried by an independent and reliable Kunpeng computing platform.
In Guangxi, the first Kunpeng industrial ecological cloud project in the region-"Magnificent Yulin government cloud (Kunpengyun)" was recently launched, which is the embodiment of the city's comprehensive promotion and application of Guangxi's digital government integration platform, and its first municipal public data open management method in Guangxi is closely related to Kunpeng's big data solution;
Looking at Zhejiang, in the process of Kunpeng's ecological settlement in Zhejiang, Zhejiang promoted the formation of the concept of "using Kunpeng", and the city was regarded as the "testing ground" of Kunpeng's ecology. Government cloud chose Kunpeng as the computing foundation, and the solutions and applications based on Kunpeng's technical architecture have been widely used in government service scenarios.
In general, it seems that the big data gameplay that only has high technology and cannot produce practical value is no longer feasible. Kunpeng's big data solution has crossed the gap between technology and application and has been implemented in many industries, scenarios and enterprises.
Among the three giants in the telecom industry, China Mobile has commercialized Kunpeng's big data solution, China Telecom has built Tianyiyun based on Kunpeng, and China Unicom has built Tiangong IT system based on Kunpeng. In terms of government affairs, Kunpeng has appeared in government cloud, such as Beijing, Guangdong, Jiangsu, Zhejiang and Guangxi, and has become the preferred technical route at present; In the financial industry, Kun Peng is helping the banking system to speed up the localization process.
It can be said that Kunpeng's big data solution has effectively promoted the development of digital economy in China, especially the application and innovation of information technology.
Under the leading edge, Kun Peng paved the way for the landing of big data technology from three angles.
1, Go More Steady —— Dealing with Important Security Issues in the Digital Age
Therefore, Kunpeng's big data solution has been increasing investment in security. The most typical thing is to ensure security at the underlying hardware level rather than the software level-Kunpeng 920 processor has a built-in hardware accelerator, and the industry is the first to support the acceleration of state secret algorithms. This kind of acceleration module built in CPU is called "endogenous security", which is technically safer with state secret algorithm.
Unlike the usual big data solutions that have to hand over more performance to monitor the system operation at any time in order to ensure security, Huawei Kunpeng's endogenous security method in big data solutions has achieved a business performance loss of less than 5% caused by encryption-not only solving the security pain point problem, but also solving the pain point problem caused by "for security needs".
2. Go more smoothly-protect the existing data software and hardware investment with compatibility.
As mentioned above, there is a very important compatibility problem between the government cloud big data solution and the existing servers. Since Kunpeng's big data solution is based on Kunpeng's processor, and the original software and hardware investments of many government and enterprise institutions are based on X86 architecture, Kunpeng needs to be technically compatible with X86 at the deployment level, so that the technology can be applied smoothly, which can also protect existing government and enterprise customers.
It can be seen that at present, Kunpeng's solution has supported the mixed deployment of big data component Taishan server and other architecture servers.
Take Jiangsu Telecom as an example. Last July, it announced the successful launch of the world's first operator big data platform based on Kunpeng processor. As a core business system, this big data platform is based on Huawei Taishan server, constructed by Kunpeng processor and open source Hadoop software, and carries the running data, storage and analysis of all production systems of Jiangsu Telecom:
During the project, the two parties cooperated to complete the compilation of open source Hadoop source code based on Kunpeng processor, successfully deployed and operated key big data business components on Huawei Taishan server, and realized the integrated deployment of traditional architecture server and Taishan server on the original cluster.
This compatibility method effectively combines the characteristics and future evolution trend of Jiangsu Telecom's big data business, gives full play to the performance of Kunpeng processor, and improves the efficiency of data storage, calculation and other resources.
3. Go wider-ecological openness can make big data have endogenous motivation.
Huawei, the main promoter of Kunpeng Ecology, has always emphasized the concept of "open hardware, open source software and empowering partners", which has also been applied in big data solutions.
For example, in Kunpeng's full-stack solution, the top-level big data platform supports Huawei's self-developed FusionInsight big data platform and open source Apache, open source HDP/CDH and ring big data platforms, which can effectively meet the needs of various scenarios. In August this year, Xinghuan Technology released a joint software and hardware solution for big data platform based on Kunpeng. Xinghuan Technology's TDH big data platform provides excellent functions in software, while Kunpeng chip provides powerful performance in hardware, extreme performance, smooth migration, rich scene support and rapid deployment, creating value for the industry.
In addition, the data virtualization engine openLooKeng led by Kunpeng is open source, which supports massive analysis across data formats, data sources and data centers, and ultimately helps the performance of the solution to be greatly improved. For example, the typical Beiming data asset management platform V4.0, based on openLooKeng technology, solves the pain points of data asset management such as complex data, different standards and difficult management, and protects and excavates data value for enterprises.
An open ecosystem will help more partners to develop computing products such as servers and PCs, help build a high-quality basic software ecosystem, and enable more eco-partners to acquire the full-scenario development capabilities of the end, edge and cloud, and ultimately promote the prosperity of Kunpeng's computing ecosystem and accelerate the application innovation of the big data industry.
After laying a good foundation and setting a good benchmark, Kun Peng entered the cycle of "the strong will always be strong"
After bridging the gap between technology and application, Kunpeng's big data solution has more and more government and enterprise practices, most of which are typical customers in the industry, which is itself a quality case in the development of big data.
Kun Peng, who has these customer resources, has actually entered the development cycle of "the strong will always be strong", not only because he has won many benchmarking cooperation cases and is ahead of the industry, but more importantly, after the channel from technology to application practice is opened, practice will continue to feed back technology and help Kun Peng hone his own technical ability, thus forming an effective positive feedback cycle.
Once this cycle is formed, big data solutions will enter the "flywheel" development process. The faster it turns, the harder it will be to stop, and it will be difficult for latecomers to catch up, and it will gradually become the most competitive choice for government and enterprise customers.
In addition, big data services have never existed in isolation. In the era of diversified computing needs, the main promoter of Kunpeng's computing industry ecology is Huawei's capabilities and ecological layout in the Internet of Things, 5G and AI. It will undoubtedly help Kunpeng's big data solution to have more imagination space for horizontal technology connection and integration, and meet the potential innovative business needs of more government and enterprise customers.
All in all, the era of data and computing is accelerating under the guidance of high-quality cases represented by Kunpeng's big data solution. In the end, "new computing power" will promote the comprehensive landing of "new infrastructure" and bring "new economy" kinetic energy, and more government and enterprise customers will enjoy the value dividend brought by technology.
* The pictures in this article are all from the Internet.
end
Zeng Xiangling
1 titanium media, Pintu, Everyone is a product manager and many other top ten authors of venture capital and technology websites;
2 Tiger Roar Award judges;
3 Writers: Best-selling authors such as business opportunities under the new normal of mobile internet plus;
Four special commentators from nearly ten newspapers and magazines such as China Business News, Business, Business Review, Sales and Market;
5 Nearly 80 columnists including Titanium Media, 36kr, Tiger Sniff, Interface, The Paper, etc.
The concept of "brain artist" (brain craftsman) has evolved into "self-media" and become an industry;
Tencent Media has sent honorary tutors and communication consultants from many scientific and technological information companies.