Cloud Network Management for Heterogeneous Datacenter Groups
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current network management systems are limited to managing single datacenter environments and face challenges in scaling to manage multiple tenant networks distributed across various datacenters, including on-premises and virtual datacenters in the public cloud, requiring a more comprehensive and cloud-based solution for unified management.
Innovation Solution
A cloud-based network management and monitoring system that deploys service instances in the public cloud for each datacenter group, utilizing a container cluster like Kubernetes, with microservices for policy management, network flow monitoring, and threat monitoring, along with local managers at each datacenter and multi-tenant services for high-level operations, ensuring isolation and efficient communication across datacenters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a network management system is deployed to manage multiple tenant networks across distributed datacenters, then the system's adaptability and management scope are improved, but the system complexity and deployment difficulty increase
Solution Approach 1:
The network management system is segmented into multiple independent service instances, each managing a specific datacenter group. Each service instance is further divided into microservices that can be independently deployed and scaled. This segmentation allows the system to handle multiple tenant networks across distributed datacenters while maintaining manageable complexity through modular architecture.
Solution Approach 2:
The system implements multi-tenant services that provide universal functionality across all datacenter groups. These services include identity management, policy enforcement, and monitoring capabilities that can be shared across multiple service instances. This universality reduces overall system complexity by avoiding duplication of common functions while maintaining the ability to manage diverse tenant networks.
2Adaptability or versatility
If service instances are deployed in the public cloud for each datacenter group, then the system's scalability and cloud integration are improved, but the deployment complexity and resource requirements increase
Solution Approach 1:
The system deploys separate service instances in the public cloud for each datacenter group, with each instance containing only the microservices needed for that specific group. This segmentation allows selective deployment of service instances based on actual needs, reducing unnecessary resource requirements while maintaining cloud integration capabilities. Service instances can be independently scaled or removed as datacenter groups are added or retired.
Solution Approach 2:
The system uses automated service instance deployment that pre-configures necessary cloud resources, networking, and security policies before actual service activation. Templates and configuration files are prepared in advance, allowing rapid provisioning of new service instances when datacenter groups are added, thereby reducing deployment complexity despite increased cloud integration requirements.
3Ease of repair
If microservices are used for different network management functions, then the system's modularity and maintainability are improved, but the communication overhead and integration complexity increase
Solution Approach 1:
The system introduces service meshes and API gateways as intermediary layers between microservices. These intermediaries handle service-to-service communication, authentication, and policy enforcement, thereby simplifying the integration complexity between microservices. The intermediaries provide standardized communication protocols and patterns, making it easier to maintain and update individual microservices without affecting the entire system.
4Speed
If local managers are deployed at each datacenter to interact with network elements, then the system's responsiveness and local control are improved, but the overall system complexity and communication requirements increase
Solution Approach 1:
The system segments the management architecture by deploying local manager components at each datacenter that handle local network element control, while centralized service instances in the cloud manage higher-level policies and coordination. This segmentation enables fast local responses to network events while maintaining overall system coherence through standardized communication interfaces between local and centralized components.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
Some embodiments provide a method for a network management service operating in a first public cloud. The method receives a definition of a group of datacenters to manage. The group of datacenters includes (i) a virtual datacenter implemented in a second public cloud for an entity and (ii) a physical on-premises datacenter of the entity. The method receives a network policy configuration defining a logical network spanning the group of datacenters. The method provides logical network configuration data to local network managers at each of the datacenters at the group of datacenters.