Since 2020, cloud-native technologies represented by containers, Kubernetes, and microservices have begun to subvert the industry’s perception of cloud computing technologies and industries since 2006. In 2019, the industry was still in the stage of hot discussion about cloud native technology, and mainstream manufacturers launched cloud native products and platforms. However, neither the industry nor users expected that cloud native technology would accelerate into the stage of large-scale applications. By the end of 2020, the trend that cloud native will disrupt cloud computing is already very obvious.
On November 30, 2020, HUAWEI CLOUD and Forrester released the white paper "Embracing Cloud Native Priority Strategy-Building Modernized Enterprise IT Infrastructure Centered on "Applications" (hereinafter referred to as the white paper). The white paper interviewed 316 enterprise-level cloud computing platform development and operation-related technical managers in July this year. The survey showed that more than half of the interviewed companies realized that traditional cloud computing and infrastructure technologies could not make enterprises more effective. Responding to changing customer needs, challenges from many aspects, and meeting the endogenous needs of ensuring business continuity and improving operational efficiency. And cloud native is the answer.
(Huawei Cloud CTO Zhang Yuxin releases cloud native white paper)
Forrester pointed out that cloud native technology will effectively release the dividends of cloud computing technology. On November 30th, Zhang Yuxin, CTO of Huawei Cloud, pointed out at the "Cloud Native 2.0, Accelerating Enterprise Intelligent Upgrade" activity: The cloud native 1.0 era is mainly for the Internet industry; and the cloud native 2.0 era is for all enterprises. In the cloud-native 2.0 era, government and enterprises will develop new cloud-native applications based on cloud-native technologies on the basis of traditional applications on the cloud, truly releasing the dividends of cloud computing technology. On the basis of the "Cloud Native 2.0" proposed in September of this year, Huawei Cloud has further launched the cloud native 2.0 panorama .
For the intelligent upgrade of enterprises, it can be said that Huawei Cloud's full-stack cloud native platform is a strategic platform that helps enterprises become a new cloud native enterprise in the cloud native 2.0 era. According to the report "PRC SDC Software Tracker, 2020H1" just released by IDC, Huawei Cloud's container infrastructure software market share ranks first in China , which makes the cloud native 2.0 proposed by Huawei Cloud more confident.
Cloud native challenges
At the Huawei Full Connect Conference 2020 in September this year, Huawei Cloud proposed "Cloud Native 2.0". Compared with Cloud Native 2.0, what problems have cloud natives encountered since the rise of Docker in 2015 after five years of development? The Forrester white paper points out that Chinese enterprise cloud platforms have encountered double challenges when applying the previous generation of cloud native technologies: the incompleteness of traditional cloud native solutions in terms of architecture and ecology hinders the modernization of enterprise cloud platforms; and cloud native open source technology The complexity and immaturity of the company bring various risks of independent research and development.
First look at the first challenge, the previous generation of cloud-native solutions is incomplete. Early cloud native solutions focused more on the container deployment, orchestration and governance of stateless applications, and used containers and orchestration technologies to improve application delivery efficiency. However, with the continuous expansion and deepening of the application scope of cloud native technology in various enterprises, the core business of the enterprise In the process of system cloud-native upgrade, the containerization requirements for distributed middleware, databases, and stateful applications, as well as the containerization requirements for emerging businesses such as AI, big data, and edge, plus the unified management and unified operation of the above applications Traditional solutions can no longer meet the demands of maintenance and unified scheduling.
Simply understand, stateless applications mainly refer to Internet applications, that is, there is no need to retain data after each execution; stateful applications mainly refer to enterprise-level applications that need to retain data for each execution for monitoring, auditing or subsequent execution. Cloud-native technology mainly solves the pain points of Internet companies, and when it comes to the cloud native 2.0 stage, it is necessary to solve the pain points of enterprise applications, including a large number of traditional enterprises. Therefore, similarly, early cloud-native solutions were mainly oriented to a single public cloud or private cloud environment, and lacked multi-cloud, cloud-side collaborative orchestration capabilities. Enterprises need a hybrid cloud architecture and meet the requirements of disaster tolerance and business continuity.
Of course, the cloud-native stage from 2015 to 2020 is a stage accompanied by the simultaneous development of emerging technologies such as big data and artificial intelligence. Therefore, the cloud-native technology system mainly provides basic platform capabilities such as general containers, and lacks the integration of emerging technology frameworks. Support capabilities, and the separation of enterprise cloud-native platforms and underlying infrastructures makes it impossible for cloud-native solutions to perform dynamic resource allocation and job scheduling based on the characteristics of business applications, resulting in waste of resources, especially AI, big data, HPC, etc. The concurrent execution of tasks greatly reduces the cost-effectiveness of resource utilization.
Next, look at the second challenge, the complexity and immaturity of cloud native open source technology. Forrester pointed out that if companies adopt open source cloud-native technologies and develop cloud-native platforms and infrastructure on their own, they will face the following risks: different components including mainstream cloud-native frameworks will iterate rapidly, and the maturity of components of each technology route will vary. Brings stability risks in independent development; the cloud native technology system covers a wide range of fields, and there are multiple routes of open source components in different fields, which bring companies the choice of technical solutions and technical sustainability risks; each open source technology component has different The complexity and dependency between components increase the scalability risk of enterprises building cloud-native platforms independently; some open source components have not yet experienced the verification of large-scale enterprise-level applications, and lack support for enterprise-level features. Potential risks in the coming years; cloud-native open source components of different development routes have brought higher learning costs and R&D thresholds for enterprise platform developers.
In response to the two major challenges of cloud-native technology in the past, Forrester pointed out: For most companies, when considering the use of cloud-native technology to accelerate the process of enterprise digital transformation, application and architecture modernization, they need to choose a clear product strategy and mature platform products. As well as partners with deep enterprise service experience, reduce the enterprise's risk in technology selection and decision-making, and focus on business needs.
The cloud native 2.0 that stands without breaking and organically collaborates
Entering 2020, cloud-native technology has begun to spread from the Internet industry to traditional government and enterprises. How to make good use of cloud-native technology in the digital transformation of government and enterprises, especially the large number of complex and technically not yet enterprise-level cloud natives that have emerged in the early open source community Technology and how to transition from the current state of enterprise IT to using cloud native technology to coordinate the overall enterprise IT architecture has become a key topic of Cloud Native 2.0.
At the Huawei Full Connect Conference 2020, Huawei Cloud CTO Zhang Yuxin proposed "Cloud Native 2.0". At that time, Zhang Yuxin pointed out that in the era of cloud native 1.0, it was mainly a "cloud native ON infrastructure" architecture, that is, a layer of cloud native technology was superimposed on traditional cloud infrastructure. In this way, heterogeneous cloud infrastructure can be managed uniformly through cloud native technology, but cloud native technology does not penetrate into every layer of cloud infrastructure. The "cloud native IN infrastructure" proposed by HUAWEI CLOUD implements cloud nativeization at every layer of cloud infrastructure, which fundamentally solves the "gap" between traditional government and enterprise transitions to cloud native. "Cloud native", the final solution is also "organic synergy" that allows traditional IT to coexist with cloud native and "organic synergy" with technologies such as big data, artificial intelligence, and cloud native.
As pointed out in the Forrester white paper, facing the multi-level needs of enterprises, cloud-native platforms need to have full-stack integration support capabilities. For example, in terms of deployment modes, they need to support multi-cloud hybrid deployment and public cloud and private cloud hybrid deployment. In terms of architecture, it is necessary to support containers, virtual machines, bare metal servers and even function computing. In terms of computing scenarios, support for cloud data centers, distributed computing scenarios, edge computing, and heterogeneous computing is required. The cloud native 2.0 proposed by HUAWEI CLOUD not only fully upgrades the original infrastructure to cloud native infrastructure, provides more efficient resources, but also improves enterprises' agile application development, delivery, operation and maintenance capabilities, and uses massive data and The efficient AI platform makes the business of enterprises smarter, while also providing end-to-end comprehensive enterprise-level security service capabilities to ensure the safe operation of new cloud-native enterprises.
At the "Cloud Native 2.0, Accelerate Enterprise Intelligent Upgrade" event, Fang Guowei, Vice President of Huawei Cloud Products, introduced the panoramic view of Huawei Cloud Native 2.0, which is a digital transformation of government and enterprise, from traditional IT and cloud computing to cloud native technology development, based on cloud native Technology development truly provides a complete technology stack and path based on cloud-based applications.
(Fang Guowei, Vice President of Huawei Cloud Products, introduces the panorama of Huawei Cloud Native 2.0)
First of all, at the cloud-native infrastructure layer, HUAWEI CLOUD puts forward the concept of “application-centric” cloud-native infrastructure, which supports multiple computing power and meets the personalized computing power requirements of different application scenarios. The bottom layer of HUAWEI CLOUD DynaSky architecture provides cloud native computing power with extreme performance for applications through software and hardware collaboration capabilities. For example: DynaSky's unique container offloading technology has greatly improved the resource utilization and performance of container clusters, which can save 30% of resource costs and increase business processing efficiency by 2 times; network pass-through technology reduces performance loss caused by software forwarding , Improve the performance of the entire network by 40%.
In the multi-cloud and edge-cloud collaborative management, Kubernetes, KubeEdge, KubeFed, Volcano, Istio, etc. are used to build an efficient, reliable, and cross-cloud unified business platform for enterprises to provide a consistent multi-cloud management experience. Among them, Volcano is worth mentioning. Batch container computing engine and Yaoguang Smart Cloud Brain. Volcano makes up for the shortcomings of Kubernetes in scenarios such as batch computing task scheduling and orchestration for AI, big data, high-performance computing, etc.; while Yaoguang Smart Cloud Brain uses global scheduling capabilities to open up the boundaries between cloud, edge and end. The delay circle as low as 5 milliseconds can meet very low delay requirements such as autonomous driving, AR/VR, and industrial Internet.
On the infrastructure layer and the multi-cloud and edge-cloud collaborative management layer, HUAWEI CLOUD provides a complete cloud container engine CCE, cloud container instance CCI, container mirror warehouse SWR, cloud container security service CGS, multi-cloud container platform MCP, and intelligent edge platform IEF, cloud native service center OSC, application service grid ASM and other application architectures, so as to achieve "container-centric unified computing, application-centric infrastructure".
Secondly, in the cloud-native application enablement layer, Huawei Cloud provides data enablement services DAYU, database GaussDB, intelligent data lake FusionInsight, one under the three themes of "business intelligence", "application agility", and "security and trustworthiness". Stand-alone AI development platform ModelArts, knowledge computing solutions, full-scenario microservices ServiceStage, fusion integration ROMA Connect, cloud native DevSecOps DevCloud, cloud middleware DCS/DMS, function FunctionGraph, data encryption service DEW, web firewall WAF, situational awareness SA , Traffic cleaning Anti-DDos and 70+ compliance certification, etc. It is worth mentioning that these cloud services and products are both traditional IT and cloud-native methods, each of which can satisfy the transition of enterprises from traditional IT to cloud-native.
Third, HUAWEI CLOUD has a large cloud-native 2.0 customer base, ranging from Internet companies to companies in finance, genetics, manufacturing, energy, and real estate, as well as extensive participation in cloud-native open source communities such as CFCF, enriching it from the perspective of industry recognition and industry practice. The cloud native 2.0 panorama of Huawei Cloud.
Zhang Yuxin emphasized that the difference between Huawei Cloud's cloud native and other vendors is "one vertical and one horizontal": "one vertical" is a full-stack technology optimization, that is, the cloud-native technology is extended to full-stack software and hardware collaboration, and At each layer, it is "standing and not breaking" taking into account traditional IT and cloud native technology; "one horizontal" is to support the development needs of various enterprises and various new business formats, and provide a unified architecture of public cloud, hybrid cloud, and edge cloud. It can realize cloud native in all scenarios with the same end, edge, network, and cloud.
New Cloud Native Enterprise Huawei
For Huawei Cloud’s cloud native 2.0, its first and most important customer is Huawei IT, which is Huawei itself. Huawei's own digital transformation has always been talked about by the industry, while Huawei's cloud-native transformation has been less disclosed before. At the "Cloud Native 2.0, Accelerate Enterprise Intelligent Upgrade" event, Su Liqing, chief digital transformation officer of Huawei Cloud, introduced Huawei's own cloud native transformation in detail.
(Huawei Cloud Chief Digital Transformation Officer Su Liqing introduces Huawei Cloud Native Practice)
Su Liqing emphasized that it is through Huawei's cloud-native transformation that Huawei Cloud realizes that an enterprise's past cannot be lost, and that while acquiring new capabilities, it must also inherit traditional capabilities. The combination of these two capabilities is the digitalization of the enterprise. The fundamental goal of transformation. Through Huawei's cloud-native transformation, it has established that the core of Huawei's cloud-native practice is to be independent and unbreakable. The new capabilities insist on being born in the cloud and being better than the cloud, and organically synergize with existing capabilities to achieve standing and unbreakable. Huawei's digital transformation practice has witnessed the development process from a "non-cloud native enterprise" to a "new cloud native enterprise".
Su Liqing introduced the eight major practices of Huawei's own cloud native transformation: end-to-end R&D operations are based on HUAWEI CLOUD, all of which are cloud-native, greatly improving efficiency; LAAS (Lab as a Service), agile realization of the "digital twin" of laboratory operations; R&D Desktop cloud, personal work codes are fully uploaded to the cloud, without feeling safe and reliable; through R&D work on the cloud, LAAS, and desktop cloud applications, Huawei Cloud R&D services more than 150,000+ R&D personnel, realizing ultra-large-scale enterprise cloud-native R&D applications , Refactoring the R&D operation model; Huawei Consumer Cloud, supporting 670 million+ users and 18 billion+ developers in 200+ countries and regions around the world, realizing "ultra-large-scale cloud-native applications"; Huawei Mall VMALL, building enterprises based on cloud-native architecture Vertical e-commerce, extremely flexible to support high concurrency such as double 11 seconds killing; smart stores, provide standardized services on the cloud, support rapid global store opening; cloud-native cloud-based exhibitions, respond to the epidemic, and achieve "customer relationship without cooling down, business without interruption" ;Huawei OWS global operator Web service, an intelligent operation and maintenance platform carried on the cloud; Huawei Cloud uses ROMA+ModelArts to realize the use of a large number of AI applications on the cloud native, such as financial reimbursement automation, contract signature authentication, site delivery Smart acceptance, etc.
It is worth mentioning that HUAWEI CLOUD DevCloud is a DevOps practice born in the cloud and better than the cloud. Since the establishment of Huawei Cloud DevCloud, the scale of software, team management, and the complexity of communication between personnel have risen sharply. Through engineering practices such as cloudification, microservices, containerization, and assembly line automation, as well as management practices such as agile, DevOps, and full-function teams, while the overall scale is increasing, version compilation, version construction success rate, system regression testing, and research and development work time Indexes such as, resource reuse rate have not been reduced, but have been greatly improved. It is the best organization and engineering practice to support cloud native architecture. Embedded in DevCloud is Huawei's best R&D model for more than 30 years, from traditional IPD to cloud-native DevOps, from ToB to ToC. In order to help enterprises in all walks of life, especially non-cloud native traditional enterprises, realize cloud native transformation and migration, HUAWEI CLOUD DevCloud adheres to a compatible and open technical framework and supports full-process DevSecOps. Today, Huawei Cloud DevCloud has served 25,000 enterprises, 70w+ developers, covering 32 software parks.
Hu Weiqi, vice president of Huawei Cloud China, said at the "Cloud Native 2.0, Accelerating Enterprise Intelligent Upgrade" event that cloud native technology represents a kind of advanced productivity. As long as it is a company that is more concerned about R&D efficiency or software version development efficiency, it will Focus on cloud native technology. Although cloud-native technology originated from the Internet industry, it is not just an exclusive technology of the Internet industry. All industries that want to improve resource elastic expansion capabilities and improve overall R&D efficiency need cloud-native technology. As for the Internet industry, previously relying on business model innovation, demographic dividends, and capital to advance, it has now entered the stage of refining technology and seeking benefits from technology. Cloud native technology will in turn affect the next development of the Internet industry. .
(Hu Weiqi, Vice President of Huawei Cloud China, second from right)
In order to further accelerate the prosperity of the cloud native industry, Huawei Cloud launched the Cloud Native 2.0 Action Plan around joint innovation, industry promotion, community contribution, and comprehensive empowerment. HUAWEI CLOUD will focus on building cloud-native communication platforms such as the "Creation Conference", continuing to contribute to the open source community, and comprehensively empowering technologies through various methods such as the Cloud Native 2.0 Industry Summit to promote the development of the cloud native industry.
In summary: HUAWEI CLOUD has focused on cloud native technologies from a very early time, and actively participated in and contributed to the open source community. HUAWEI CLOUD not only provides full-stack cloud-native products and services. Every cloud-native product or service at each layer is based on Huawei’s perspective, which combines traditional IT and cloud-native technologies to "stand without breaking", helping companies to follow suit smoothly. Transition from traditional IT to cloud native. The “organic synergy” of technologies such as artificial intelligence and big data will allow companies to develop cloud-native applications such as the Internet, the Internet of Vehicles, the Internet of Things, 5G, and AI under cloud-native technologies, thereby becoming a true new cloud-native enterprise and releasing The long-term dividend of cloud computing technology. (Text/Ningchuan)