Share Via
Introduction
As Nile set out to offer wired and wireless networks built around a service, a big part of the offering was to own the entire device lifecycle process for each of the network elements deployed at customer sites. In addition to ensuring that a customer is running the best network hardware possible with a deterministic design at guaranteed performance levels, managing software upgrades for the network elements within the Nile Service Block is a big part of the service.
Traditionally, upgrading software on network devices has been a complex and potentially disruptive process, often leading to unexpected changes and significant costs in terms of time and resources.
Nile introduces a service-oriented approach to software upgrades, ensuring that the network remains at or exceeds its previous performance levels after each upgrade. We recognize the sensitivity of software upgrades, especially in legacy environments, and strive to minimize potential disruptions. If you would like to catch the demo before we dive deeper, here it is:
Nile Services Cloud and the Digital Twin
Our approach starts with bringing determinism to the process, applying closed loop automation and leveraging the Digital Twin within the Nile Services Cloud of each network to perform checks that ensure the goodness of a software upgrade.
Because we’re involved in each step of deploying a network we leverage AI and the data from Day -1 when site surveys are performed, as well as capturing operational data to perform this closed loop operation. Each of the following steps are discussed and how they play a role:
- Setting maintenance windows
- Using the Nile Digital Twin
- Pre and post upgrade sanity checks
Maintenance Windows
As part of Nile’s standard operational workflow, we work with customers to put together recurring maintenance windows that are suitable to them. We then take responsibility for performing automated software upgrades during these customer provided maintenance windows. Customers have the option to provide restricted windows, which means maintenance operations including software upgrades cannot be performed during these times.
Digital Twin and Orchestrated Upgrades
Once a software upgrade begins, our automated process orchestrates a procedure that leverages the Digital Twin. The following list outlines the 360-degree view of how the data for a customer’s Nile Access Service implementation is captured.
1. A pre-upgrade snapshot of the network is taken. This includes several key data points related to the Nile Service Block itself, such as, the number of active access points, sensors, switches, links, and other useful network telemetry such as LLDP information within the Nile elements. This can include links speeds, interface counters and more.
2. Using data from Nile physical and virtual sensors we also capture access point visibility and RF neighborhood information.
The information from bullets #1 and #2 are what we refer to as inside-out metrics that are related to the elements themselves. These metrics are deep instrumented telemetry from each network element within the Nile Service Block, that allows for better cognitive decision making based on near real-time system resource data like memory utilization, CPU statistics, as well as interface and buffer counters. Better data, better outcomes.
3. With the above information in hand, Nile adds to our differentiation by also taking a snapshot of the end devices before the upgrade. These are devices on the network that are static in nature like printers, wired phones and similar IoT devices. These typically are devices that the network sees as always connected and do not exhibit roaming characteristics. Because these devices are always connected, they are good candidates to use when measuring the goodness of the upgrade.
The expectation is for these devices to consistently remain present and establish connections without encountering issues. They offer an outside-in approach, completing a 360-degree view once endpoint device connectivity is restored, ensuring that everything functions seamlessly, just as it did before the upgrade.
Post upgrade checks
Digital Twin within the Nile Services Cloud will receive telemetry from the Nile elements after the upgrade completes. If the longstanding devices reconnect after the upgrade completes, telemetry related to those are also received. A post upgrade snapshot is then taken which is compared to the pre-upgrade state automatically by the Digital Twin.
Based on an acceptable level of discrepancy, the AI-enabled cognitive decision module then decides to either declare the upgrade as successful or automatically revert the software version to the earlier one if the discrepancies in the pre and post snapshots are widespread. Essentially, checking that the longstanding devices perform as expected, this further instills confidence in the goodness of the software upgrade.
Conclusion
Nile ensures the highest level of software upgrade quality by taking a 360 degree approach based on AI networking principles to ensure the state of the network is on par or better after the upgrade based on quality data and metrics. The key is to not only leverage data related to the core Nile Service Block but also endpoint device connectivity, and pre and post snapshot comparisons. This unique capability makes it possible for the Nile Access Service to completely automate and eliminate traditionally manual operational burden in maintaining software releases across enterprise wired and wireless network infrastructure.
