Why Nokia Migrated to Automated Global Data Centre Networks

Share this article
Share this article
Prioritise Us on Google
Nokia's data centre fabric is a network of switches that work together to provide infrastructure for connecting traditional and AI-based applications installed on servers (Credit: Nokia)
Several factors drove Nokia to pursue a network migration - one of them was the network suffering repeated outages that could last up to nine hours

In Nokia's factories, every hour of network outage cost US$100,000.

That was when their data centres experienced frequent outages, that could last up to nine hours at a time, nearing million-dollar losses by the network, essentially a waste of employees' and the company's time.

Futurum and Nokia's report on its global data centre network migration reveals that if a network lost a heartbeat for just two seconds in its old environment, the database would go down and take over two hours to recover.

Significant costs caused by these network failures and outdated operations led Nokia to transform its legacy infrastructure by consolidating onto a new automated modern data centre fabric solution.

Nokia's data centre networks are now built on Nokia data centre switches, running their SR Linux network operating system (NOS) managed by the Nokia Event-Driven Automation (EDA) platform.

Through modernising their network operations, the company reduced challenges its data centres faced and achieved drastic improvements in reliability.

(Credit: Nokia)

Transforming outdated technology

Nokia's team adopted its own data centre fabric solution and data centre switching hardware, and deployed an automated EDA Software-as-a-Service (SaaS) delivery model. 

One issue for employees working in the data centres was operational overhead associated with hosting and maintaining the management platform, as one leader remarked that upgrading the server and managing infrastructure would be a painful process. 

The SaaS deployment allows the team to focus on network design and automation workflows rather than on maintaining the underlying platform and facing complexity.

Mitch Ashley, VP and Practice Lead of Software Lifecycle Engineering at The Futurum Group, writes on Nokia's blog: "Nokia wanted to create an environment where network changes could be made without hesitation, validated before deployment and carried out by teams who trusted both the system and the process.

Mitch Ashley, VP and Practice Lead of Software Lifecycle Engineering at Futurum Group (Credit: Futurum)

"That meant shifting to an architecture built around reliable automation, intent-based configuration and integrated change validation.

"And it meant designing for scale from day one – so that lessons learned in a single site could be carried forward globally."

Gradual deployment by phased migration

Nokia IT carried out the deployment of their new automated technology in their data centre sites by phases. They started by implementing the new fabric in one site, particularly the Europe dual data centre sites which were built on legacy vendor infrastructure.

After these trials went through successfully, the team adopted the same approach to the US data centre sites with trained staff for oversight.

Futurum and Nokia's report states that the brownfield migration process resulted in zero unplanned outages caused by the new network, which may have been due to all configurations being validated in the automation platform's digital twin before being applied.

Youtube Placeholder

Benefits of the migration

As a result of the initial deployment and eventual network migrations at the Europe and US data centres, Nokia saw benefits to its choice to modernise its data centre network.

This included an 80% reduction in network-related incident reports at initial deployment. This overall means that less time is spent by engineers and IT staff on troubleshooting and huge cost savings.

It also eliminated chronic outages by old manufacturing applications, including one that was used for 20 years, which reportedly occurred almost every month on the legacy network. Migrating to the SR Linux and EDA-based fabric meant that these disruptions vanished.

The automation of network operations also means that day-to-day-tasks can now be handled via high-level intent definitions which leaves specialised tasks to It engineers. 

Nokia IT believes that once the new system is fully rolled out, a small team will be able to operate all their data centres globally.

The migration shows that complex, legacy-laden networks can be transformed into simple cloud-era infrastructure, and can be used as a blueprint for other large enterprises to reap the benefits of automated systems and result in cost-savings. 

Company portals

Executives

  • Mitch Ashley

    Vice President and Practice Lead for Software Lifecycle Engineering