HUAWEI CLOUD Cloud-Native Infrastructure Ushers In the Cloud Native 2.0 Era
Oct 16, 2020
On September 24, 2020, Zhang Yuxin, CTO of HUAWEI CLOUD, delivered a keynote speech titled "Technical Innovations Bridge Divides and Accelerate Intelligent Upgrade" at the 5th HUAWEI CONNECT. In his speech, Mr. Zhang officially released the HUAWEI CLOUD Cloud-Native Infrastructure Solution, which marks the start of Huawei's exploration in the Cloud Native 2.0 era.
Mr. Zhang said, "Cloud-native technologies bring in agility and efficiency unparalleled by other technologies and are the cornerstone of future-oriented digital infrastructure. In Cloud Native 1.0, cloud-native technologies have accelerated application development and simplified O&M. However, cloud-native technologies focus on the application layer and are simply overlaid with cloud infrastructure to deliver the "Cloud Native on Infrastructure" architecture. In this architecture, the compute, network, and storage resources are unaware of the requirements of applications for high availability, high performance, and auto scaling, and cannot support services across clusters, regions, or clouds. Therefore, enterprise services and applications cannot implement cloud native–based transformation."
The Cloud-Native Infrastructure Solution released by HUAWEI CLOUD originally proposes "Cloud Native in Infrastructure". This converged architecture marks the beginning of the Cloud Native 2.0 era and serves as a cornerstone of digital transformation for enterprises.
Redefining infrastructure: HUAWEI CLOUD implements application-centric resource scheduling based on the QingTian architecture. It uses software-hardware synergy technologies to provide enterprises with cloud-native infrastructure that delivers the optimal performance, cost-effectiveness, and experience. For example, the innovative container offloading technology based on the QingTian architecture greatly improves the container cluster resource utilization and reduces the resource costs by 30%. The intelligent computing scheduling system performs computing power decision-making and scheduling, dynamic overcommitment, and time-based reuse based on runtime metrics (application traffic, node resource usage, network topology, etc) and SLA requirements (performance, cost, reliability, etc). With the network passthrough and elastic network interface (ENI) technologies, QoS-guaranteed network resources can be directly and, more importantly, dynamically allocated to containers, resulting in no performance loss and a two-fold improvement in bandwidth. In addition, cloud-native storage pools allow storage resources to be quickly and automatically allocated, scaled, and migrated based on container requirements. This provides secure, reliable, and high-performance data management capabilities for stateful applications.
Re-enabling ubiquitous applications: Enterprises are facing challenges with managing resources across clusters, regions, and clouds. Based on technologies such as cloud-native cluster federation and edge-cloud synergy, HUAWEI CLOUD has developed a multi-cloud and edge-cloud synergy management solution to help enterprises build an efficient, reliable, and cross-cloud service platform for a consistent management experience across clouds. In terms of applications, enterprises can enjoy rapid image distribution, application scheduling, load balancing, and one-click fault locating across clouds, thereby developing services and simplifying service O&M in multi-cloud environments. In terms of network and storage resources, multi-cloud hyper-converged storage pools and cross-cloud network model optimization are used to enable data scaling and migration, and service SLA assurance across clouds. In this way, the application and infrastructure layers are streamlined. In addition, HUAWEI CLOUD allows enterprises to extend their services to the edge with Intelligent EdgeFabric (IEF), a powerful edge service that provides industry-leading technologies such as an edge-cloud synergy computing framework, offline autonomy, network anti-jitter, and fault self-healing.
Refactoring application architecture: Stateless applications are still dominant in many enterprises using clouds. Such monoculture constrains the speed and scope of cloud native implementation. To encourage service diversity in enterprises, HUAWEI CLOUD builds Operator Service Center (OSC), a cloud-native application center that supports more than 130 cloud-native applications, including stateless applications in the microservice architecture, stateful applications (such as distributed middleware, AI, and big data services), and emerging applications (such as blockchain, edge, and 5G services). With OSC, enterprises can build their own cloud-native service architecture and achieve "all-in-container". HUAWEI CLOUD is committed to building an open and standard application ecosystem. Based on cloud-native operators, HUAWEI CLOUD standardizes the lifecycle management of distributed middleware and builds a unified platform for enterprise application distribution, rollout, and O&M. The Istio-based Application Service Mesh (ASM) provides non-intrusive microservice governance, which is an advanced, unified application governance service. Volcano, a batch scheduling engine integrated with mainstream computing frameworks, provides efficient scheduling to support AI, big data, offline, and real-time computing services.
HUAWEI CLOUD: Advancing with Clear Goals
Core contributor to key technologies: Huawei is the only Chinese enterprise among the founders of Cloud Native Computing Foundation (CNCF) and also the first platinum member of CNCF in Asia. Huawei has led the development of a large number of core subprojects and key features such as networking, scheduling, and multi-cloud technologies. Huawei ranks first in Asia in committed code, and accelerates the development of cloud-native technologies.
Implementer of cloud-native technologies: HUAWEI CLOUD container services have made remarkable achievements over the past five years. By the first half of 2020, more than 30 million containers had been provisioned and used internally in Huawei, including Huawei BP IT, consumer cloud, R&D cloud, EI services, and WeLink. Externally, HUAWEI CLOUD containers have been deployed in multiple industries, such as Internet, finance, automobile, biomedicine, and energy, building repeatable business models for other industries that are about to embrace cloud native technologies. According to the IDC report, HUAWEI CLOUD ranked first among Chinese cloud service providers in terms of the market share of container software in 2019. Based on the technical accumulation in the implementation of cloud native in these industries, HUAWEI CLOUD contributes KubeEdge (a cloud-native intelligent edge computing project) and Volcano (a batch scheduling project) to CNCF, accelerating the adoption of cloud native in edge, AI, and big data services.
Leader of industry development: HUAWEI CLOUD has been committed to promoting and developing a cloud-native ecosystem. At HUAWEI CONNECT 2020, HUAWEI CLOUD shared its practices and experience in developing and implementing cloud-native technologies in different industries and enterprises. At the HC Cloud Native CXO Roundtable, HUAWEI CLOUD, together with 12 partner enterprises, officially launched the Enterprise Cloud Native Construction Guide White Papers project, and released the subsequent white paper plan to work with analysts and industry organizations. Through continuous output and sharing, HUAWEI CLOUD will help more enterprises understand and go cloud native.
Fang Pu, General Manager of HUAWEI CLOUD Container Service Domain, said, "Cloud-native technologies have matured and become the cornerstone of enterprise digitalization. HUAWEI CLOUD has helped thousands of enterprises in more than ten industries successfully implement cloud-native technologies. In the Cloud Native 2.0 era, HUAWEI CLOUD will continue to innovatively employ cloud-native technologies to more industry scenarios, and transform cloud-native infrastructure from resource-centric to application-centric. This will help more enterprises complete digital transformation and upgrade their service architecture."