Accelerating Network Automation
and Innovation with Middleware

networks and includes myriad varieties of native xponders, disaggregated xponders, and OTN switches, all via the same set of APIs. As a result, their middleware becomes the optical domain controller communicating bi-directionally with the network southbound and northbound (so to speak) with the customer’s home-grown orchestration tools. An additional benefit is provided by real-time inventory for both networks, once again via a set of vendor and technology-agnostic APIs.

Rapid Integration. Continuing down the path of APIs and their relationship with middleware reflects other advantages. A single API, by function, can also mean little to no development, or delays in the development, associated with a new solution for the physical network. In many instances, this new transport technology accelerates this progression, creating even faster savings. Moreover, given the persistence and strata of unpredictable supply chains, this can also facilitate timeline capacity by actually increasing vendor choices when bottlenecks occur, and delivery is delayed by an incumbent vendor.

In this case, one of the nation’s largest providers of shared communication infrastructure with over 40,000 cell towers and approximately 85,000 route miles of fiber supporting small cells and fiber solutions uses a 5G backhaul switch in two different POP sites connected to filters and amplifiers. In turn, these are connected point-to-point across the outside fiber plant with up to five passive add/drop filters, each tuned to a set of eight wavelengths that feed cell tower routes. In terms of middleware, both multiple vendors and multiple technologies are combined into a single view of middleware for inventory as well as automated maintenance. In an effort to eliminate enrollment time and human error, network designs are fed directly into this solution from an external inventory system. The result? Validation happens at no cost to the CSP in the lab of the multi-vendor middleware, while promoting competitive advantage—and, keenly, disrupting optical transport innovation.

Single Data Structure and AI-Enabled Tools. By using multi-vendor middleware, data is normalized to facilitate AI-centric tools that are implemented between layers of the physical network, including: Layer 0 physical, Layer 0 passive, Layer 1 transponder, Layer 1 line, and Layer 1 functions embedded in the Layer 2 devices, and even extending to some Layer 2 functions.

In certain instances, there are dependencies, (e.g., where products of one tool are essential to the function of another tool). Having a common database and single APIs allow elegant automation as visibility to passive devices and logical inventory of their presence and function make AI tools sharper and more precise.

A primary focus for this single data structure is the experience of one of the largest telecommunication companies by revenue and a leading provider of mobile telephone services in the U.S.—and whose main focus in its largest US tier 1 is maintenance. DWDM, disaggregated Circuit Emulation, SONET, and TDM networks are all interconnected to visualize the true end-to-end circuit paths through multiple network segments.

Key automated maintenance functions include Circuit Analysis Tool (CAT) and Proactive Optical Wavelength Restoration (POWR). In this instance, both are vendor and technology agnostic, available through the UI and API, and capable of gathering real-time data from every network port in its service path. This enables myriad outcomes, to include seamless integration to ticketing systems, test platforms, and external customer “self-test” capabilities. In terms of metrics, CAT API is executed on roughly 5,000 services a day from TDM through DWDM networks. POWR API is focused exclusively on isolating troubles and even autonomously clearing issues on next gen OTN networks executed nearly 40 times a day.

In its next phase, maintenance automation evolves into service layer alarm correlation for Circuit Emulation networks. This correlation engine identifies an events root cause, correlates children services and alarms, and only reports on the common lowest level service. As a result, this can reduce NOC ticket volume by 90%, while also reducing Mean Time To Repair (MTTR). Coupled with a middleware solution, these combined tools provide automated troubleshooting, trouble isolation, and trouble resolution. In other words, an obligatory single pane of glass for users, and automation for any layer 0/1 network across the globe.