E.ON expands its shared computing environment with Cisco HyperFlex Systems and Cisco ACI.
With a focus on smart power grids and innovative customer solutions, international energy firm E.ON is simultaneously expanding and consolidating its shared computing and networking infrastructure with Cisco HyperFlex and Cisco ACI.
Industry: Energy and utilities
Location: Essen, Germany
It was 2019 when E.ON, a leading European energy conglomerate, started connecting its energy grids and transforming its operations. While its four subsidiaries in Germany—Schleswig-Holstein Netz AG, Bayernwerk Netz GmbH, Avacon Netz GmbH, and E.DIS Netz GmbH—remain independent, they now share technology resources and responsibilities to maximize operational and cost efficiency.
"We're aligning our OT [operational technology] systems across the country and splitting the workload between the subsidiaries," Benjamin Bubbers, network administrator for Schleswig-Holstein Netz AG, said at the time. "We want one combined effort instead of four individual, repetitive efforts."
The goal has been achieved, he reports, with a shared infrastructure built on the proverbial shoulders of Cisco HyperFlex Systems and Cisco ACI. Spread across four regional data centers, the hyperconverged compute and storage systems are managed by Bubbers' four-person team at Schleswig-Holstein Netz AG, and the software-defined network that ties them all together is managed by their counterparts at Bayernwerk Netz GmbH.
"So far, so good," Bubbers says of centrally managing distributed infrastructure for four independent companies. "We have excellent integration between compute and network systems, and we've implemented our internal processes and handoffs. To be honest, getting everyone aligned from a process standpoint has been harder than technology implementation and management."
The renowned simplicity and scalability of Cisco HyperFlex, key factors in E.ON's decision to adopt the platform, have already been put to the test. The company's original deployment of four stretched clusters has been expanded to ten, and Bubbers hasn't needed to bolster his team to manage the additional hardware.
"Managing 10 clusters is the same as managing four of them," Bubbers says, noting the use of Cisco Intersight Virtual Appliance to manage the 84 nodes contained within the clusters. "We love Cisco HyperFlex."
“Managing 10 clusters is the same as managing four. We love HyperFlex.”Benjamin Bubbers, Network Administrator
They also love the improved density. With the company's new Cisco HyperFlex clusters supporting twice as many virtual machines per node compared to their initial deployment, E.ON has significantly improved resource and cost efficiency.
"The environment is being simultaneously expanded and consolidated," Bubbers says. "We're getting the same performance with fewer nodes, and less infrastructure means less cost."
In addition to scaling the shared computing environment, Bubbers and his team worked with their peers at Bayernwerk Netz GmbH to establish tight integrations between Cisco HyperFlex, Cisco ACI, and VMware vCenter.
"Network and infrastructure teams now have full visibility of endpoint groups, VLANs, bridge domains, and virtual machines, all of which can be easily moved from one cluster to another," Bubbers says. "When we have to refresh or update a cluster, we simply move everything to another cluster to avoid downtime."
The company has also established three isolated network zones to further improve infrastructure resiliency and data security. The "orange zone" is closest to the internet and contains routers, proxy systems, and gateways. The "yellow zone" contains the energy conglomerate's primary workloads. And the "green zone" is reserved for databases and systems that require the highest levels of security.
"Our workloads and endpoints are separated and secured based on priority and level of sensitivity," Bubbers says. "Orange can't talk to green, and vice versa."
With its compute and network systems now fully integrated, E.ON is preparing to deploy an eleventh Cisco HyperFlex cluster for new, containerized workloads. Bubbers' team will provide the infrastructure, memory, and CPUs, and the subsidiary companies will manage the virtual machines on their own.
"It's a great help for them because they don't have to worry about buying, deploying, and managing servers and storage," Bubbers says. "They can get it all from us and remain focused on their workloads and business needs."