Business Challenge
Our client, a large global Service Provider was looking to automate their core network from day 0 to day N and onwards. This included the onboarding of new devices such as routers, switches and the optical layer and then provisioning new services on top of the network based on the requirements of both internal and external customers. It was mandatory that the interface (API) should conform to industry standards and the system should function without needing manual intervention where ever possible.
In partnership with the customer’s software development team, Flint’s orchestration and automation team developed the essential service models required to making the automated vision a reality.
Initial Challenges
- The customer’s network is a true multivendor environment with interchangeable components, meaning that the development process for new services for different configuration styles of different network operating systems had to be done in parallel.
- To ensure feature parity for all vendors the Flint team set up an extensive CI/CD pipeline for running automated tests on virtual devices of different types.
- While on paper the different vendors’ equipment supported the same or equivalent features, it often turned out that the reality was somewhat different requiring Flint to help communicate the deficiencies to the vendors and find suitable work around.
Solution Implementation Process
The process comprised of 3 phases. A bottom-up approach was used to create the required components.
Phase 1:
- A translation layer was built that transforms the static (service) configuration to network configuration. This is implemented using Cisco NSO and in practical terms means a number of smaller services exist that can be used to manually piece together a complete working network
Phase 2:
- The translation layer was augmented to enable it to react to changes in the network
- All services from phase 1 were augmented with the logic to examine their health on the network and report this over a standardized interface to the orchestrator.
- The system is then able to react to events and either take direct action or notify others. Actions can be simple, like use an out of band management network to reach an unreach- able device, or quite complex, like notice a device needs RMA, order a replacement, wait for the field engineer to replace it and finally restore the services on the new device.
Phase 3:
The focus was on scalability: automatin of capacity planning was done, entailing monitoring the performance of the network and automatically requesting new (physical or virtual) resources to keep in line with the agreed SLAs.
Customer Benefits
- Creation of on-demand instances of Layer 3 Virtual Private Networks (VPNs) by internal and external customers in minutes, rather than days or weeks
- Quick integration with any OSS/BSS systems on the customer side
- Reduction of time and cost associated with new service introduction
- Eliminating the potential for manual error creating a significantly better quality of service and customer experience
Are you facing a similar challenge, or would you like to learn more about our services? We’ll be happy to hear from you!
