Introduction: Cloud network management is an intelligent network management operation and maintenance platform based on ali Cloud network’s years of technology and experience. It provides enterprises with the ability to manage operation and maintenance of network life cycle, making deployment faster, operation and maintenance more efficient and network more transparent.

1. The background

Cloud network management is an intelligent network management operation and maintenance platform based on ali Cloud network’s years of technology and experience. It provides enterprises with the ability to manage operation and maintenance throughout the network life cycle, making deployment faster, operation and maintenance more efficient and network more transparent.

1.1 User pain points

Difficult installation and deployment

Most traditional OSS are deployed locally. Users are responsible for physical resource planning, middleware installation, and installation package deployment. Professional network and IT maintenance personnel usually need several days or even weeks to deploy the OSS.

Difficulty in centralized management

The vendor NMSS of network devices can only provide local Web access, but cannot be managed in a centralized manner, let alone across vendors, and cannot provide users with a unified management perspective. Therefore, multiple NMSS are required to switch between them.

Network Expansion Difficulty

With the expansion of services, offices and retail outlets often need to set up shop in different parts of the country. Currently, network engineers log in to a console port for network delivery, which is inefficient and error-prone.

2. Product introduction

2.1 Scope of network management

The following figure defines the full life cycle of network management. Traditional network management systems mainly manage the resources and operation and maintenance (O&M) of online nes during the service phase of the network. However, they cannot manage the full life cycle of network devices.

Figure 1 Network management lifecycle

The NETWORK management of the CLOUD network management system runs through the whole life cycle. When the network is not online, the network architecture can be planned and designed offline. At the time of construction and delivery, the defined network architecture is implemented in the way of project, and the whole delivery process is controllable with manageable quality. After the acceptance, the network officially enters the service stage, during which the core demands of the network are stable operation, fewer failures, fast location and fault recovery after failures occur. The monitoring, change, inspection and other modules of the CLOUD management ensure the stable operation of the network in the service stage until the network goes offline. Throughout the network life cycle, resource management ensures that network resources are consistent with the real network.

2.2 Product Functions

2.2.1 Construction delivery

  • Architecture design

Before the network is online, you can plan the network in advance. The network architecture defines the interconnection specifications and technical specifications of the network, provides graphical architecture design and management functions, flexibly arranges the network topology in graphical mode, and configures specifications of each network component.

In the figure, the network architecture is designed in a visual way, and the number and connection relationship of network modules are defined.

Figure 2 Network architecture design

A standard configuration file that can be imported with one click is automatically generated based on the designed network planning and configuration specifications.

Figure 3 configuration file generation

  • The construction of the delivery

Construction delivery is to deliver the network scheme to equipment in the form of project, and control the delivery process and guarantee quality in the form of work order in each project.

FIG. 4 Construction delivery project implementation

2.2.2 Intelligent O&M

  • Situational awareness

The whole state of the network can be sensed in real time through the global view and branch view.

Health: Scores the network status based on device alarm status, network inspection status, and monitoring coverage.

Dynamic topology: The LLDP and MAC scan technologies are used to update the quasi-real-time topology and display the device connection relationship and port information in real time.

  • Integrated monitoring

A network of office branches or stores usually consists of multiple layers and devices in various network forms. These devices include the gateways that connect to carriers, network devices (routers or switches) that forward data, wireless devices that provide WLAN, and various terminal applications. A monitoring scheme covering the whole link from the network to the end is necessary, because often the anomalies of the end may be the problem of the wired network. The failure of the wired side may affect the use of multiple terminals or applications at the downstream end.

For devices at different network levels, such as wired networks, wireless networks, and terminal applications, the CLOUD management system monitors the network running status from the gateway to terminals through various monitoring technologies, such as active collection, event reception, plug-in deployment, and active dial-up. In the following monitoring view, you can monitor wired and wireless indicators, such as the traffic of the upper interface on the switch, the traffic of the egress to the carrier, the number of AP terminal connections, and terminal monitoring information in one view.

  • Network layout

The CLOUD management system provides a visual process orchestration engine. Users can drag and drop atomic capabilities of network operations to ensure transaction integrity and security during service delivery.

  1. Complete business process choreography definition

2. Visualized delivery of a single-step configuration process

  • Fault self-healing

Based on the network choreography capability of the CLOUD management system (CSS), alarms can be processed together in daily high-frequency fault scenarios to quickly stop faults and recover services.

3. Architecture introduction

3.1 Technical Architecture

  • Protocol plug-ins

The cloud management protocol plug-in integrates SSH, Telnet, Netconf, SNMP, and GRPC protocols required by the management network to manage common commercial network devices.

The protocol plug-in communicates with the cloud collection and control instance through the encrypted secure channel to efficiently execute the device operation instructions and collection tasks sent from the cloud, and compresses the data to the cloud in real time for analysis and display.

The Agent of the protocol plug-in can be output in various ways, such as its own hardware, software installation package and integrated SDWAN gateway.

  • Acquisition control

The acquisition and control instance is deployed in cloud and plays a connecting role in the overall framework of cloud network management.

Scheduling engine workflow:

1. Receives the tasks delivered by the network choreographer and decomposes and schedules the tasks based on task priorities and scheduling policies. 2. Translate commands into specific commands for operating devices based on the model of the device manufacturer.

3. Send the command to the Agent to receive the execution result

4. Task execution results analysis and assembly.

Template management: Templates are classified into user templates and device templates. User templates have service meanings and are directly referenced in network choreography, ignoring vendor differences. Device templates are detailed to vendor and model granularity, and vary from vendor to vendor. For example, in the device template of ACL policy, Cisco and Huawei have different commands.

  • Network layout

Network orchestration is responsible for the uniform abstraction and definition of business model. Network orchestration connects atomic capabilities according to business processes to form concrete network schemes. In the network scheme, the acquisition and control instances are driven by the flow engine to perform atomic operations in each step, and the transaction integrity and link tracing are controlled by the way of work order.

  • application

Network orchestration provides capabilities and data interfaces for applications in the form of apis. The application layer implements specific capabilities such as resource management, Network inspection, network change, and fault recovery, and supports O&M personnel in daily network o&M and service configuration.

3.2 Deployment Architecture

The cloud management system is deployed as a SaaS, and the cloud management system instance is opened at the minute level. On the user side, only the probe is deployed (the hardware version requires only power-on and network reachable).

The computing and storage resources of the CLOUD management are ali cloud resources, which can be expanded at any time according to the specifications. The CI/CD function realized based on ALI Cloud ASK cluster can be iteratively launched at any time.

4. To summarize

Cloud Network Management is committed to creating a SaaS network operation and maintenance management platform free of deployment, easy to use and centralized management for complex, heterogeneous and multi-branch offline networks.

The original link to this article is the original content of Ali Cloud, shall not be reproduced without permission.