Minimizing runtime trouble in broadband network automation
It’s time to talk some more about automation in broadband networks. Last time we discussed Altiplano automation and low-code development benefits in design time. Now we turn to the runtime benefits.
Scripting doesn’t leave much of a trace
Scripting the interaction between a lot of different elements is a challenge. The power of integrated policy and workflow engines, of course, is in the ability to program for a whole range of network scenarios without writing a single line of code. But even when you feel you have all the right automation scripts in place, it’s not until runtime that you get to see if they fly or fail.
Scripting doesn’t leave much of a trace. If something does go wrong and you need to make a change, you typically don’t know where to start. You won’t know what was executed exactly, and you’ll likely need to reflow, because of missing or incorrect information. There is generally no easy recipe when a script fails somewhere in the middle of execution and you want to recover the situation. Analysis and reconstruction will take time and effort. At some point, you wonder if it was the good-old ‘fat finger’ misstep or some form of human error, where we press the wrong key or put the wrong data in the wrong field.
Automation in runtime
We’ve packed powerful features into our solutions that overcome these issues. When an onsite engineer makes a mistake or when the node configuration is left in an undesirable state, Altiplano can detect the misalignment and suggest corrective actions to restore the node or service. Because every change is tracked and versioned, you can go back to any previous version with a click of a button. Altiplano can also pre-validate configurations and constraints before they are active on the live network, which avoids incorrect settings being pushed to the network. Devices can be pre-provisioned in the cloud without the nodes being available or hooked up to the network. Whenever the onsite engineer completes the physical installation and boots the node, the node will call home and synchronize with the latest configuration data in the cloud. This decouples the subscriber provisioning from the actual installation of physical devices on-site.
The health of each subscriber’s service is automatically monitored based on alarm and state data. Massive telemetry data gets collected in time-series databases and data is available when you need it. The collected data can also be reviewed over time. If something goes down over night, for example, node data is available even if the node is offline. Or in the case of a pesky intermittent fault, even if a service is currently healthy, you can consult historical health data and still discover the cause.
Tackle complex network automation scenarios
We’ve successfully automated card swaps (swapping a failed card for a new one); node replacements (swapping a failed node for a new one, or a new type); customers taking their ONTs to a new location; fiber moves due to infrastructure changes; subscriber upgrades; technology migrations from GPON to XGS-PON; and updates of existing services and profiles. Our profile manager application lets you make profile changes without being an expert in JSON files or having to recompile any code.
A big challenge for operators is keeping all subscriber configurations aligned over time. Which configuration is active on which subscriber? Today’s networking tools execute ad-hoc in fire-and-forget mode and are state unaware as they lack a feedback loop from the network. An intent-based tool like we have in Altiplano is much more effective as all actions are declarative and can take into account that the realization can change over time. It ensures network tasks be maintained and repeated. You get feedback from the network, obtain good visibility into what is happening and can better plan ahead.
Learn more how our Altiplano Developer Portal makes networks easier to program, operate and customize by operators and developers.