Nokia AIMS: a transformation in warehouse inventory management

Warehouses are the hidden heart of the economy, and without them operating efficiently, our global supply chain would grind to a halt.

Warehouse activities are increasingly automated, but many logistics operators still depend on manual counts to ensure accurate inventory. Manual counts are costly, slow and prone to human error.

Nokia Bell Labs saw an opportunity to tackle these challenges and bring clarity to warehousing. As Nokia’s industrial research lab, its mission is to innovate with purpose, pursuing responsible, sustainable technologies that will have a demonstrable impact on society.

Nokia developed a solution, Autonomous Inventory Monitoring Service (AIMS), which used AI computer vision to navigate autonomous flying drones around warehouse spaces, scanning inventory barcodes to create a ‘digital twin’ of the facility. 

The drones could scan 7x-10x faster than humans, operating when scheduled and keeping inventory counts accurate. Nokia estimated AIMS would deliver warehouse operators 40% ROI over three years.

Pushing the limits of edge Kubernetes

To make a cutting-edge solution like AIMS work depends not just on advanced hardware, but on a complex supporting software stack deployed on a bare metal edge server inside the warehouse.

This edge server communicated with each drone to synthesize collected data into the warehouse’s digital twin and for data analytics.

Nokia’s engineers had built the AIMS edge solution on the foundation of Canonical’s Ubuntu Pro and K8s using containerized applications. The build process was manual, conducted server by server through a traditional CLI. Ongoing operational tasks for each Kubernetes-based software stack were performed manually.

This approach enabled the Nokia AIMS team to rapidly prove the AIMS concept — but as customer interest grew, Nokia knew it needed a way to automate and scale rapidly across hundreds of warehouse sites.

Sustainable scale with Spectro Cloud Palette

After a thorough evaluation process, Nokia chose to partner with Spectro Cloud and use its Palette Edge platform to manage the lifecycle of its AIMS edge Kubernetes clusters.

With Palette, Nokia was able to:

• Model the full software stack for deployment to each site, including its choice of Ubuntu Pro OS, K3s lightweight Kubernetes distribution, plus integrations for observability, networking, CPU operator, certificate manager and much more.
• Bootstrap bare metal servers and consistently deploy this software stack, repeatedly and at scale.
• Effectively deploy to ‘airgapped’ locations, thanks to Palette’s local content bundles, which ensure that clusters can access the container images they need even if they can’t reach the internet.
• Centralize management of connected clusters with an intuitive NOC-style UI, and automate day 2 operations such as patches and upgrades.

Along the way, Nokia had access to award-winning solution support from Spectro Cloud, covering the full software stack.

Clear results

Nokia proved the AIMS solution with Palette in production with customers.

For Nokia’s engineers, the provisioning process had gone from one day to one hour — a huge improvement in efficiency and a reduction in repetitive manual tasks. Onboarding new customers and new sites was much easier than before.

Ultimately, Palette enabled Nokia AIMS to deliver a better customer experience, with fast zero-touch provisioning in their warehouse environments and a ‘hardware as a service’ experience of the AIMS product.