This article is "CSDN cloud" to native applications platform and so on Ali, head of tert same interview, Alibaba middleware authorized reprint. In this article you will learn about cloud computing technology development path container era, aspects Ali cloud product matrix, iterative technology and future trends in the field of container and container technology Evolution Ali cloud.
To the container as the representative of native cloud technology, the shortest path to the era of cloud cloud release value
"In the past we often virtualized cloud platform as an interface and interact with customers, for enterprises flexibility also brings certain management complexity; emergence of container virtualization on the basis of the package up a layer, and gradually become one cloud platform to interact with customers new interface, build the application, distribution and delivery to standardize on this level, dramatically reducing the enterprise IT implementation and operation and maintenance costs and enhance the efficiency of business innovation.
From the dimensions of technology development, the open source cloud computing become more standardized so that the container has become the standard application distribution and delivery, can be applied with the underlying operating environment decoupling; Kubernetes become a standard resource scheduling and choreography, shielding the bottom differences in architecture, help applications run smoothly on different infrastructures; on the basis of the establishment of the upper layer application abstraction such as micro-services and service grids, and gradually form a standard modern evolution of application architecture, developers only need to focus on their own business logic, without paying attention to the underlying implementation, native cloud is passing methodologies, tools and ideas to reshape the entire software stack and technology life cycle.
To the container as the representative of cloud native technology, using open, standard technology system to help companies and developers to build and run elastically expand in the cloud, fault tolerance is good, easy to manage, easy to observe the system has become released cloud value the shortest path. "In reference to the evolution of container Tert - calendar with notes. The earliest creation and application container technology is the Internet company, today with open standards cloud native ecology, so that the container technology spread rapidly, more and more companies and developers use containers to build applications together to enjoy this technology bonus.
What is the current enterprise customers demand for new container technology? T with this said, cloud native application landing process, the security needs of business users is one of the most concern. Traditional RunC container with host Linux kernel sharing, providing limited by CGroup and namespace isolation, as more and more enterprise customers concerned about container security, in recent years new high isolation, began to run when security, including MircoVM (Kata Container, FireCracker) direction and gVisor security sandbox direction.
Ali cloud and gold suit ant teamwork, introduction of container security sandbox technology, in September 2019 issued a security sandbox based RunV lightweight virtualization technology, compared to RunC containers, each container has RunV separate cores, even if container belongs kernel is compromised, it will not affect other containers. Ali cloud container service provides end to end native cloud security architecture, including the infrastructure, security, supply chain security software and run-time security, providing all-round, three-dimensional, multi-layered security protection.
The second demand is the hybrid cloud infrastructure, the cloud has become a trend, a lot of customers for business reasons to consider hybrid cloud of way, infrastructure and security architecture differences will cause fragmentation of the cloud and enterprise IT operation and maintenance system, increase management complexity. In the era of cloud native to the container, Kubernetes represented shielding technology infrastructure differences, as a base to promote the application-centric hybrid cloud architecture arrival 2.0 to meet the demands of users.
While enterprises in operational efficiency, R & D efficiency, operating cost and system fault tolerance, maintainability, also put forward higher requirements. Ali cloud in the whole container product development process, to address business pain points, although individual businesses to the cloud has many different demands, but the container and native cloud as a pervasive technology, different levels to meet different business needs.
Turning Ali cloud for innovative container technology and products, t with stressed, "Ali cloud want to continue walking routes community in the cloud native areas, fully compatible and open source technology that uses the scale of complexity Ali economies and Ali cloud customer scenarios rich degree of innovation, the formation of the leading native cloud technology and community feedback to build standard. in early 2019 we will be landing mass production of native cloud best practices to settle down, open to the community open source projects OpenKruise way, on the one hand to help enterprise customers in the process of exploring cloud native, the detours reduce technology fragmentation, while promoting upstream communities gradually improve and enrich the application of automated management capabilities; in October 2019, Ali goes with Microsoft jointly released an open application model open application model (OAM ), OAM is focused on the standards described in the application lifecycle, can help application developers, application operation and maintenance personnel and infrastructure operations teams to better collaborate. in this model, the developer responsible for defining application composition, dependence and architecture ; application operation and maintenance personnel responsible for running custom applications Configuration and operation and maintenance requirements, such as publishing strategies and monitoring indicators and infrastructure operation and maintenance team can be deployed to different application environments, customized configuration parameters, this separation of concerns the design, you can apply the definition, operation and maintenance capabilities and infrastructure decoupling, make application delivery more efficient, reliable and automated to address the industry pain points. Serverless by a higher level of abstraction that allows developers to eliminating resource management, operation and maintenance and other routine work, so that simplify development, extreme elastic, pay-by with Ali cloud computing products to create a function FunctionCompute, providing programming event-driven; providing Serverless application hosting platform for use of SAE, users only need to provide applications to achieve, while the platform is responsible for elasticity, automated operation and maintenance; provide ECI facing the container Serverless products and Serverless Kubernetes, on the other hand is also pushing some of the Serverless of traditional techniques, such as databases and messaging middleware. driven by technological innovation, Ali cloud cloud native wants to be the best product realization, continue cloud largest of play Native application practice, the most abundant native cloud family of products, the largest native cloud customer groups, the most comprehensive cloud-native open source contribution of advantages, service a wider range of enterprise customers and developers. "
Cloud architecture evolution and upgrade of 11 double eleven years, creating the best technological innovations soil container
"Ali cloud and other cloud vendors biggest difference is that Alibaba's core business operations in the cloud, the best form of innovation soil, that is our most advanced technology, will first try in Alibaba own business system, been large-scale use, universal and prove the value of its technology and then open to the customer. "Turning to the progress of the vessel Ali cloud, with emphasis on tertiary any technology Alibaba will get to try their own business systems and applications, 2011, Ali cloud began to move the container door, 2013 Docker advent of rapid integration Ali cloud container its advanced concepts, and promote the overall evolution of the Group's container business in 2015, and the development and evolution of this series are in fact not from open double 11 big promotion needs, such as a container full of 11 large double to help promote rapid elastic expansion.
11 double experience in the activities of the millions of containers supporting dual 11 activities carried out smoothly. Because of the very large scale business so that its complexity is very high, which also brought greater challenges for container technology. For example, in a container mirroring the distribution process, a release distributed tens of thousands of images, such huge traffic is not a small challenge. To achieve ultimate efficiency requirements, Ali cloud using P2P technology to achieve rapid distribution of large-scale high-volume, achieved within 10 seconds to complete the download mirrors across the room to start container.
Container technology for significantly affect bis 11 further comprising specific technical embodiment mix section, tert same represented by the mixing section, which are within Elibaba group-wide can save IT costs about 30%, it is possible bis 11 this specific time period where the cost per million transactions declined by more than 75%.
Container, micro-services, future trends of artificial intelligence: collaboration, integration
Container technology has been widely recognized by the industry, the prospects for the future development depends not only on its excellence in the field of technology, but also needs more technology integration, making it a successful product of technological progress with the times. The main run of the stateless early Kubernetes Web applications, such as micro-services-based applications Apache Dubbo / Spring Cloud, and now more and more enterprises core business data and business intelligence innovative business also runs on Kubernetes. Ali cloud to cloud their products, for example, including enterprise-class distributed application services EDAS, real-time computing platform Flink, elasticity AI algorithms service EAS, block chain BaaS platforms are deployed on Ali cloud services Kubernetes ACK.
From the application architecture evolution, the development of the container to promote the development of micro-services. Micro service in early landing a big problem encountered is inconsistent architecture splits caused by the complexity of the operation and maintenance and the environment, it is through the vessel and its operating environment for application packaging and isolation, a good solution to the pain points of the micro-service system, so that micro service is developing rapidly. The introduction of micro-services architecture to solve some of the problems at the same time, the invasion of the R & D framework, the framework of the iterative development and iterative coupled, and not friendly enough to support multi-language environment, also in the management more complex. Therefore, the community began to try to Service Mesh, gradually sinking ability of micro service from the frame capacity of platform capabilities, we can see the container with the micro-services and promote each other.
"Cloud native and AI is the perfect partner, the two are mutually energized." T with emphasis on integration of AI in reference to the container. First is the emerging field of AI, not so much historical baggage on architecture, in addition to the requirements of AI calculation itself has been resilient, resource efficiency and deployment of efficiency, container technology can solve these problems. GPU, FPGA, proprietary ASIC chip and other new infrastructure, huge count to increase, it also brings the difficulty of management and maintenance, the use of Kubernetes provide unified management of heterogeneous resources and efficient scheduling, enhance the elasticity support fine-grained sharing, you can improve resource utilization 3-5 times. AI for container native cloud technology is also a huge help, AI often represent business scenarios for how this native cloud technology is more pervasive universal verification provides a wealth of space, so as to enhance the maturity of native cloud technology.
Container technology has been around for more than six years, Kubernetes rapid development is not news, but that does not mean that container technology ecosystem has developed smoothly. In contrast, the container and its surrounding technical system is still maintained rapid development. Turning to the future focus on new technologies, new direction, with t-admitted to let go of all container environment, not only the traditional IDC, but also to go to a public cloud, private cloud, edge nodes, networking, big data, AI and other scenes in the hope that the use of native cloud technology to reduce the threshold for the use of cloud computing, cloud the real turning point.