Search
Home Search Center IP Encyclopedia Online Courses Intelligent Model Selection

What Is Network Digital Map?

As a next-generation network O&M technology, the network digital map is a digital twin of the physical network world. It acts as the digital brain of enterprises that intelligently manages the cloud, network, devices, applications, and users in a unified manner. It also helps customers to implement network O&M based on the dynamic and HD digital map instead of a static topology. Customers can therefore view the network on the intuitive network digital map, significantly improving the network O&M efficiency.

Why Is Network Digital Map Needed?

Before digital maps became the norm, we had to often rely on our own memory or physical maps and road signs to get to a destination. We may not be aware of any changes in road conditions on the route ahead, resulting in detours. The navigation map makes this a thing of the past. It turns the traditional static geographic map into a dynamic and real-time digital world. With it, we can get the latest traffic data in real time, such as congestion and road closure information, improving our travel experience.

Enterprises are accelerating their pace towards becoming digital and intelligent. As the foundation of digital infrastructure, network connectivity plays an increasingly important role in promoting digital transformation across industries. As digital transformation gathers pace, enterprise services are developing rapidly and networks are drastically growing in both scale and complexity. However, there is no corresponding increase in the number of network O&M personnel of enterprises, meaning that the O&M personnel have to do more. To add insult to injury, there is no unified view that can centrally display the enterprise network health, leading to a poor network experience, numerous complaints about network faults, and inefficient fault rectification, thereby hindering the digital transformation pace of enterprises.

What Problems Can the Network Digital Map Solve?

Unlike the conventional static network topology, Huawei's network digital map is a real-time, dynamic, and HD network map. The network digital map uses a host of key technologies such as high-speed data collection, AI, big data computing, as well as route simulation and verification algorithms to achieve multi-dimensional network visibility, path navigation, search and positioning, and deterministic application experience. It also provides abundant capabilities such as real-time visibility into network quality, fault demarcation and locating, and self-healing, helping customers evolve from static topology-based O&M to dynamic and HD digital map-based O&M. In this way, the network digital map intuitively displays the network status in real time, maximizing the network O&M efficiency.

Similar to a navigation map, the network digital map is a digital twin of the physical network. It provides a visualized, easy-to-interact, and experience-centric digital intelligent management platform. To better understand the service scenarios of the network digital map, we can compare the network digital map to a navigation map.

Comparing the network digital map to a navigation map
Comparing the network digital map to a navigation map
Table 1-1 Comparing the network digital map to a navigation map

Function

Navigation Map

Network Digital Map

Holographic visualization

  • Road map
  • Traffic heat map
  • Real-time satellite map
  • Application view
  • User view
  • Network view

Positioning

  • Placemark location
  • Congestion location
  • NE location
  • Packet loss location

Path navigation

  • Route from location A to location B
  • Network path from NE A to NE B

Key assurance

  • Dedicated bus lane
  • Priority for emergency vehicles
  • Assurance for key events
  • Assurance for key users

Service expansion

  • Ride hailing and food delivery
  • Asset and terminal management

Application of Network Digital Map in a Data Center

In the data center domain, networks used to be separated from applications and are unaware of application changes. To address the limitations, the network digital map provides a unified network and application view, making it possible to implement service-oriented navigation and positioning in the network world and expedite the digital transformation of data center networks.

Application of network digital map in a data center
Application of network digital map in a data center
  • One-map network visibility: Typically, more than 20 O&M systems coexist on a multi-cloud and multi-vendor data center network, leading to fragmented management of different network zones and separated O&M of networks and applications. Huawei's network digital map uses ecosystem integration plug-ins to standardize data structures and perform digital modeling based on multi-dimensional data such as network traffic, topologies, and configurations, improving the topology accuracy from less than 60% to greater than 99%. In addition, the map can dynamically and intuitively display the multi-dimensional mappings and forwarding paths between applications, networks, and services.
  • One-second fault demarcation: In the conventional O&M mode, services and networks are troubleshooted separately, lacking a unified view. In addition, it is costly to deploy the conventional network performance monitoring (NPM) and device-layer indicators are invisible, leading to difficult and slow troubleshooting. This cannot meet financial regulatory requirements. By leveraging the xFlow (intelligent full-flow analysis) capability, the network digital map adopts a host of flow analysis technologies such as Encapsulated Remote Switched Port Analyzer (ERSPAN), port mirroring, and In-situ Flow Information Telemetry (IFIT) to enable unified management of application and network performance indicators, hop-by-hop fault diagnosis across clouds and vendors, and raw packet backtracking and forensics, thereby building a unified O&M platform for devices, networks, and applications. All in all, this achieves a future-oriented shift from the network device-centric O&M to the application experience-centric intelligent O&M.
  • One-click network optimization: As data centers evolve from single computing power to diversified computing powers, the number of AI training tasks increases dramatically. AI training flows are often small in number but large in volume. Moreover, the conventional hash algorithm can easily cause load imbalance on networks. As a result, the effective network throughput is merely 50%. Oriented to scenarios where masses of tenants share clusters and each cluster has 1000/10,000 GPUs/NPUs, the network digital map uses the network scale load balancing 2.0 (NSLB 2.0) algorithm to optimize the network load, increasing the network-wide throughput to 98% and the AI training efficiency by 20%.

Application of Network Digital Map on a Campus Network

On a campus network, the network digital map accurately analyzes information at the network layer, user terminal layer, and application experience layer to gain insights into people, events, and things on the campus network and their relationships with the network. This facilitates network troubleshooting and service experience assurance.

Application of network digital map on a campus network
Application of network digital map on a campus network
  • Three-layer status awareness: As the number of devices and terminals increase on a campus network, maintenance work will become overwhelming for O&M personnel if they cannot gain visibility into network status. The digital map of a campus network is a digital twin of the physical network. It can detect the status of the network layer, user terminal layer, and application experience layer in real time. When a device or application is faulty, or user experience deteriorates, the digital map will display it.
  • Minute-level AI-based locating: Wireless access scenarios of campus networks are increasingly diversified. However, access failures and network interruptions occur frequently. In addition, it is technically challenging to troubleshoot wireless network faults. If not promptly detected and handled, these faults may cause user dissatisfaction or even complaints. The digital map visualizes the user status based on the user journey and displays the real-time user experience status in temporal and spatial dimensions. When user experience deteriorates, a message is displayed promptly, including information such as when it occurs, where the user is located, and to which AP the user is connected.
  • One-stop experience assurance: Audio and video services on campus networks grow rapidly. However, traditional audio and video service experience assurance is always performed manually, and complex configuration steps need to be completed on devices one by one. This process is time-consuming. Therefore, as quick experience assurance is in urgent need, the digital map is ideally suited as it can implement one-click delivery of experience assurance. You only need to select assurance objects (VIPs or applications), and configurations will be automatically generated and delivered with just a click of the mouse. This immensely boosts experience assurance efficiency.

Application of Network Digital Map on the WAN

In the WAN domain, the network digital map offers real-time visibility into network status and performance, and can make reliable judgments and decisions based on the dynamically changing network data, implementing automatic load balancing of network traffic and differentiated service assurance. All of this helps you build a digital foundation for autonomous networks, thereby accelerating the upgrade of premium private line packages for a better monetization potential.

Application of network digital map on the WAN
Application of network digital map on the WAN
  • One-map network visibility: The network digital map collects a massive amount of key network data in real time. The data includes network traffic and bandwidth utilization, link delay and packet loss rate, as well as static attribute configurations and protocol status of Layer 3 links. A digital twin engine with outstanding data processing, storage, and query capabilities is adopted to provide real-time and reliable network data via unified data modeling and diversified application data governance. This enables the real-time refresh of the digital map.
  • One-second fault demarcation: The network digital map provides fine-tuned IFIT in-band flow measurement technology that can manage network-wide link status in real time. When a link is congested, congestion data such as packet loss, delay, and jitter information can be reported to the digital map in seconds through telemetry. The network digital map then analyzes the data, quickly detects network performance changes, and displays the changes on its GUI to guide resource scheduling for rapid fault rectification. This can slash the fault locating time from several hours to minutes.
  • One-click network optimization: The network digital map supports automatic network optimization. It provides intelligent path computation for large-scale networks. That is, when link congestion occurs or service quality deteriorates, the network digital map can centrally compute paths based on the latency, bandwidth, packet loss rate, and other differentiated service level agreement (SLA) requirements of different services. In this manner, the map can obtain the paths best suited to the services' SLA requirements in real time and rapidly divert the traffic from the congested link to a new link. This eliminates manual intervention and improves network utilization by 30%.

Network Digital Map Architecture

The network digital map architecture comprises three layers: data access layer, service platform, and service application layer.

  • Service application layer: This layer provides essential functions of the network digital map to visualize network topology for customers, thereby facilitating routine O&M activities and service modifications.
  • Service platform: This layer provides crucial technical modules, which implement functions at the service application layer or ensure that necessary dependencies are met. The digital twin engine is responsible for data governance and digital modeling, and is one of the core foundations of the network digital map. The network simulation module is responsible for network simulation modeling, path computation, and breakpoint analysis, and is the technical cornerstone of network topology navigation.
  • Data access layer: This layer centrally manages various data sources and collection drivers of the network digital map service, and provides data access services for the network digital map system.
Network digital map architecture
Network digital map architecture

Key Technologies of Network Digital Map

Massive Data Collection

A traditional network management system (NMS) uses the Simple Network Management Protocol (SNMP) to obtain device indicators. However, this approach is unable to provide experience-centric O&M. First, SNMP uses the pull mode to collect data, in which the NMS requests and then the device replies. The data collector interacts with a device in request-reply mode. The device needs to reply to each request sent from the data collector. This process is resource-intensive and time-consuming, particularly in the case of a massive data query. Second, SNMP uses a rigid data structure, which requires multiple data requests to complete an effective data collection. Based on this mechanism, an SNMP data query is typically sent every five minutes. If query requests are sent too frequently, normal services of the device will be severely affected. Telemetry is used by the network digital map to remotely collect data from devices at a high speed. Devices periodically push device information to a collector, providing real-time and accurate network monitoring.

Network digital map collecting massive data using telemetry
Network digital map collecting massive data using telemetry

Telemetry organizes data based on the Yet Another Next Generation (YANG) model, encodes data in the Google protocol buffer (GPB) format, and transmits data through the Google Remote Procedure Call (gRPC) protocol. This improves data collection efficiency and facilitates intelligent interconnection. Telemetry is more than 20 times faster than SNMP in terms of massive data collection. It supports data collection at an interval of 10 seconds and has the following advantages:

  • Proactively pushes data, reducing the pressure on devices.
  • Pushes data periodically to avoid data inaccuracy caused by network delay.
  • Monitors a large number of network nodes, making up for the disadvantages of the traditional network monitoring mode.

Big Data Analytics and Machine Learning

With the development of cloud-based services, wireless terminals, and IoT, service models are undergoing major changes. Network nodes as well as service faults and fault causes increase exponentially. Traditional device-centric and passive O&M cannot meet the requirements for responses to faults in the digital era. Engineers are also unable to manually analyze massive amounts of data.

To address these issues, machines are introduced to automatically associate, analyze, and mine big data (collected to the big data warehouse at an interval of seconds) based on the fault feature library, and identify exceptions based on expert experiences. Additionally, the aggregation of massive amounts of big data creates conditions for discovering unknown correlations and causal relationships through machine learning.

Big data modeling and analysis of network digital map
Big data modeling and analysis of network digital map

By continuously learning, modeling, and analyzing massive amounts of data, machines can identify complex service models, build dynamic baselines to predict data trends, and mine unknown correlations to identify potential faults and root causes.

Multi-layer Display of Network Digital Map

The network digital map intelligently analyzes the collected data and displays the data in multiple layers and dimensions, providing intuitive visibility into network status in the digital world.

Multi-dimensional display of network digital map
Multi-dimensional display of network digital map

The bottom layer is a Geographic Information System (GIS) map. The locations of sites and network nodes are displayed on the GIS map in a unified manner. Multiple formats of online and customized maps can be loaded, such as tiles in xyz format as well as Open Geospatial Consortium (OGC) services like Web Map Service (WMS) and Web Map Tiled Service (WMTS). Offline loading is also supported. You can set the GIS coordinates of sites and NE nodes to automatically place sites and network nodes on the map based on the GIS coordinates. Multi-level topology zooming and automatic topology layout are supported. NE and link status information can be displayed based on the network topology. Service and application information can be displayed on the network, such as path navigation and end-to-end (E2E) service visibility.

AI-based Intelligent Network Topology Restoration

Conventional NMSs restore network topologies based on Link Layer Discovery Protocol (LLDP). However, this method is unable to adapt to complex heterogeneous networks. When legacy devices or third-party devices that do not support LLDP exist on the network, or LLDP is disabled for security compliance, the network topology cannot be accurately identified and restored.

AI-based intelligent network topology restoration
AI-based intelligent network topology restoration

Using the AI inference technology for multimodal representation, the network digital map can perform precise topology identification and restoration based on device configurations, ARP entries, MAC address entries, port traffic characteristics, and LLDP data.

Application Visualization and Fault Demarcation

The network digital map identifies and classifies application protocols based on Smart Application Control (SAC), and uses Packet Conservation Algorithm for Internet 2.0 (iPCA 2.0) to analyze application experience and locate faults. iPCA 2.0 measures packet loss based on color bits. To be specific, iPCA 2.0 periodically sets and resets a certain color bit in service packet headers on the ingress interface to measure packet loss based on the measurement interval.

iPCA and fault locating
iPCA and fault locating

In the preceding figure, packet loss is used as an example. In a certain period of time, the number of outgoing IP packets on port 2 is 1000, and the number of incoming IP packets on port 3 is 900. Therefore, packet loss occurs on link ACH2, and the packet loss rate reaches 10%. In this case, O&M personnel only need to rectify the fault on ACH2 based on the analysis result.

Application and VIP User Experience Assurance

A multitude of applications are deployed on enterprise networks, user locations are constantly changing, and audio and video services are increasing rapidly. Therefore, prompt application and VIP experience assurance are pressing needs. Conventional service assurance requires manual configurations on each device. The process is complex, time-consuming, costly in human resource terms, and highly dependent on expert experience, which is difficult to meet service requirements.

The network digital map allows you to configure assured applications and VIP users with just a click of the mouse on the GUI. The system automatically generates service configurations based on service intents and delivers the configurations in batches in one-click mode, vastly improving the efficiency and accuracy of key service assurance.

Application and VIP user experience assurance
Application and VIP user experience assurance

The system reserves dedicated resources for important applications and VIP users. For example, wireless APs reserve OFDMA spectrum resources for VIP users and audio and video services. With dedicated fast lanes for important applications and VIP users, issues such as meeting freezing/lagging and VIP user experience deterioration caused by swarm traffic movement and audio and video traffic bursts can be minimized.

Share link to