“Recently, the 2019 OpenInfra Global Summit was held in Shanghai. Telecom operators, service providers, application developers, and community contributors from around the world participated in the summit to jointly promote the innovation and development of open source technology.
Recently, the 2019 OpenInfra Global Summit was held in Shanghai. Telecom operators, service providers, application developers, and community contributors from around the world participated in the summit to jointly promote the innovation and development of open source technology.
The OpenInfra Shanghai Summit focused on key open source projects, such as OPNFV, Kubernetes, ONAP, Ceph, etc., as well as open source projects hosted by the OpenStack Foundation: OpenStack, StarlingX, Airship, Kata Containers and Zuul. The agenda of this summit includes ten topics, namely: 5G, network function virtualization (NFV), edge computing; container infrastructure; artificial intelligence (AI), machine learning (ML) and high-performance computing (HPC); continuing Integration/Continuous Delivery (CI/CD); Getting Started for Novices; Open Development; Private Cloud and Hybrid Cloud; Public Cloud; Security; Practice Workshop. The theme of this conference covers application cases, special reports and live demonstrations of more than 30 open source projects such as OpenStack, Kata Containers, Kubernetes, Hadoop and StarlingX. ??
Among them, edge computing is one of the protagonists of this summit. ZTE and industry partners shared the theme of “How to achieve coordination and management of?heterogeneous and multi-vendor cloud on telco edge” at this summit. Speak. ZTE said that telecom operators can provide network services, MEC computing power, IaaS, PaaS, SaaS and many other services at the edge. Edge infrastructure is the underlying support for all these upper-layer services. The edge infrastructure of telecommunications is very different from the core infrastructure. The latter will only have a small number of clouds and solutions. In contrast, the edge infrastructure is more open and diversified. The number and types of telecom edge clouds are more than those of core clouds. How to manage these heterogeneous edge clouds from multiple cloud service providers will be a problem that telecom operators need to face when they provide various clouds and services on the edge.
Specifically, it can be achieved through the following ways:
1) Realize the LCM management of heterogeneous cloud platforms: a cloud management platform/installation and deployment tools, management of virtual machine clouds and container clouds, implementation of application deployment and resource expansion.
2) Unified monitoring of heterogeneous cloud platforms: With the help of the existing OpenStack ceilometer and CNCF Prometheus and other existing cloud platform monitoring indicators, and then shield the difference in monitoring indicators of the underlying cloud resources, report and report in a generally unified format. show.
Mainly include the following functions:
vMEC edge cloud management supports the connection of heterogeneous resource pools (through community native interfaces);
vMEC edge cloud management supports image management and distribution deployment (via community native interface);
vMEC edge cloud management supports flexible resource management configuration (through community native interface);
vMEC edge cloud management supports unified user authentication;
vMEC edge cloud management supports unified operation and maintenance (through community native interfaces and interface enhancements).
The structure diagram of heterogeneous resource pool management is as follows.
On the core side, there are NFVO/VNFM/EMC/OSS, etc.; at the edge city level, the edge cloud management is deployed to realize edge resource management and operation and maintenance management; at the district/county access level, there will be heterogeneous resource pools. Due to the need to meet various different applications, different computing resources, storage resources and network resources are required to meet the requirements. So there will be different resource pools, the current mainstream ones are OpenStack, Kubernetes, VMWare, etc.
Since the resource pool is on the edge access side, everyone is currently paying more attention to its existence, and the community also has different implementations. Here we will discuss the implementation of heterogeneous resource pools.
The first is Kubernetes + Openstack. In Kubernetes based on physical machine deployment, Director manages OpenStask and Kubernetes resource pools uniformly, and provides tenants with two different types of resources, of which the physical resources are completely independent.
The second is K8S over OpenStack. Deploy Kubernetes on OpenStack, and Director manages the Kubernetes resource pool in a unified manner.
The Kubernetes cluster is deployed in a project of OpensStack, the computing & storage & network are all provided by OpenStack, and the application runs in the container POD. The advantage of this solution is that you can quickly co-sign K8S clusters, and use OpenStack’s tenant isolation and network functions.
The third type is: OpenStack over K8S. OpenStack is deployed on Kubernetes, and Director manages the OpenStask resource pool in a unified manner. The current form of existence in the community is StalingX
First, a Kubernetes cluster is deployed in the infrastructure, and then the OpenStack service runs on the Minion node of the cluster in the form of POD, and the application runs on the virtual machine created by OpenStack. The advantage of this solution is that it can quickly co-sign OpenStack, and flexibly implement resource orchestration through the Kubernetes Helm chart function.
Edge cloud management can provide edge cloud management capabilities after managing different types of resource pools for northbound orchestration and other resource applicants. Its functions include:
1) Image management
2) Authentication and authentication
3) Resource management
4) Operation and maintenance management
Edge cloud management can also realize southbound interface docking with different types of resource pools for resource management and operation and maintenance management, including:
Openstack, Kubernetes and other third-party vendor resource pools.
1) Unified certification management
2) Mirror management (upload, query, distribute, delete, etc.)
3) Computing & Storage & Network Management
4) Resource organization
5) Service Management (Chart)
6) Application management, etc.
The edge cloud management can easily realize the management of the edge resource pool, which greatly improves the efficiency of users. There are more functions to be tapped through edge cloud management, which will bring more benefits to users.
As a gold member of the OpenStack Foundation and one of the main code contributors, ZTE is committed to promoting the evolution and development of OpenStack. Edge computing is an innovation that applies cloud computing technology to edge infrastructure, and it is still in the early stages of development. With the advancement of 5G network construction and large-scale commercial deployment, various applications are accelerating, and edge computing is expected to continue to suffer for a long time. In the future, ZTE will continue to increase its investment in the edge computing open source community and actively promote the maturity of community technology.