zorandim75 - Fotolia
Total Gas & Power has replaced ageing storage hardware with about 500TB of Nutanix hyper-converged infrastructure in a move that has allowed it to slash management time, simplify IT team structure and support a move towards DevOps.
The company has saved more than £95,000 a year in maintenance, support and licensing costs, and expects further economies.
Total Gas & Power sells gas and electricity to business customers and is a UK subsidiary of French oil giant Total. It has its headquarters in Redhill, Surrey, and several offices around the UK and Europe.
Key to its operations are sales lifecycle functions built around a customer relationship management (CRM) system and supporting databases, with large amounts of data in an Oracle back end.
Technology architect Dominic Maidment said its infrastructure comprised “20 years’ worth of technology” and had typified “the bad old world” of legacy systems, with a “high degree of complexity and systems bespoke to Total Gas & Power”.
He said the idea of moving away from that was “very painful” because of dependencies on applications and operating systems. And although some of that changed with virtualisation, “there was still a lot of old hardware on the floor”, he said.
Some systems were so old, said Maidment, that “you couldn’t get spares, except out of a landfill somewhere – so we were constantly worried about a box failing or a room getting too hot”.
He added: “We were on non-scalable hardware and it felt very brittle, constantly running in crisis mode. There would only have to be one failure and ops people would be up at 3am waiting for spares from EMC, for example.”
Storage comprised HP, EMC and NetApp, which supported the company’s VMware estate.
The company has also deployed converged infrastructure, with Flexpod, the converged product of a joint venture with NetApp storage and Cisco compute and networking.
Maidment said: “We tried to get it to work, but you’d see engineers shake their heads when they were anywhere near it. A sight you don’t want to see is three people around a screen whispering and swearing.”
Total Gas & Power has now deployed 26 all-flash Nutanix nodes that support 350-400 virtual machines. These run VMware at its production and disaster recovery (DR) sites, with Nutanix’s own AHV hypervisor at branch offices.
Nutanix runs a mix of domain servers, applications and SQL servers, plus web servers and Linux systems for containers.
Read more on hyper-converged
- Can compute, storage and networks in one box be a better option than shared storage? And which organisations can most benefit from it, and for what scale of deployment?
- Hyper-converged infrastructure can be an easy win when it comes to datacentre infrastructure upgrades, but there are pitfalls. We discuss how best you can sidestep them.
Hyper-converged infrastructure combines compute and storage into single nodes that can be built out into clusters in a scale-out fashion.
For Maidment, the key benefit is “far less drama than a traditional or partially converged estate”.
“It simplifies deployment of applications and provides a single view of capacity management,” he said.
“I had always assumed you needed three people to provision anything – a storage guy to carve it out, a virtualisation guy to hook it up and a network guy to provide IPs and authorise it. Now we only need one engineer to do that, which is great for us on a small, multiskilled team.”
But Maidment’s team has not been able to deploy Nutanix for everything. He said these include “a bunch of flat file stuff, OSs – for example, Unix servers – and things that are tightly coupled to storage”.
He added that it was “difficult to know what to do with Oracle workloads as these needed to be deployed in a manner compliant with licensing”.
Meanwhile, the Flexpod still runs the company’s VDI (on Microsoft Hyper-V). “We’re coming up with a plan for that,” said Maidment. “It’s still got life on it.”
The Nutanix deployment supports the company’s move towards DevOps by hosting the toolset and application programming interfaces (APIs) on Linux Red Hat.
“Nutanix provides a standard environment, with a fast response,” said Maidment. “With it being so performant, it allows things to work quickly and on-premise. The main thing for DevOps is that it goes like the clappers.”