Kurhan - stock.adobe.com
Ireland-based gift voucher retailer Smartbox has ditched traditional server-storage architecture for Nutanix hyper-converged infrastructure with hybrid cloud capacity in a move which has seen it consolidate disparate IT in several European countries to a single datacentre.
The move to hyper-converged – where compute, storage and virtualisation hypervisor come in one node – has also seen the company able to replace several specialist IT disciplines with one engineering resource.
Smartbox sells customer experiences as gifts through 16 web brands across 10 European countries. They can be bought online or physically in boxes, and gift recipients can login to make reservations, exchange and upgrade gifts.
Smartbox’s IT operations are subject to rapid scaling upwards and downwards and depend hugely on applications built on Docker containers.
The company had grown through acquisition and inherited a massively varied infrastructure across European operations that comprised servers and storage from a range of suppliers, including Dell physical servers, VMware and IBM V7000 storage.
The key pain points with the existing infrastructure were that it suffered from fragmentation and silos tended to build up, not helped by the functional structure of the IT organisation, said director of IT operations, Paul Cash.
The company’s had a “squad” structure in which vertical devops teams were aligned to infrastructure and had their own environments, but this suffered from being siloed. “We’d find engineers from one squad taking storage and compute and that would impact other environments,” said Cash. “It was a huge necessity that we move on from that scenario. We had no roadmap at the infrastructure level, so we started to look at ways to standardise our infrastructure.”
Read more about hyper-converged infrastructure
- Simplicity, high availability, scalability and reduced cost are among the benefits of hyper-converged infrastructure. We survey the benefits, pitfalls and the key suppliers.
- Hyper-converged infrastructure vs NAS and SAN shared storage. 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?
The IT team considered all-physical servers and a VMware virtualised environment but eventually opted to go with Nutanix hyper-converged infrastructure.
Cash’s team initially tested a Nutanix node on its QA function where the team started slowly with a few threads and processes and gradually ramped up the workloads. Eventually, the company deployed 32 Nutanix hyper-converged nodes – in a deployment that includes G4, G5 and G6 models – with around 500TB of storage capacity in total.
Smartbox uses the Nutanix AHV hypervisor, and Prism Pro as its infrastructure monitoring and management tool. Containerised applications in Docker are orchestrated by Nutanix’s Kubernetes distribution, Karbon. Smartbox doesn’t retain any persistent data from its containerised applications, said Cash.
While IT operations in-house have been centralised on Nutanix at the company’s Dublin datacenter, web front-end operations and burst capacity are handled using the AWS public cloud.
The key benefits of the move to hyper-converged infrastructure centre on the simplification of IT hardware. HCI comes in nodes of server, storage and virtualisation which can connect in grid style to form clusters. So, a key benefit is that a number of legacy technological disciplines are superseded by the need to manage one simplified infrastructure.
Smartbox is therefore able to use one engineer to manage the platform where several were required before for physical and virtual servers, plus storage. “We have about 1,100 machines running virtual instances so we need to be able to automate provisioning resources automatically,” said Cash. “We get up to the minute real-time monitoring, where previously it was all spreadsheets, running reports or going to the teams.”
A key reason to invest in centralising on Nutanix was because Smartbox didn’t want a multi-datacentre strategy, said Cash. It wanted to stick with one datacentre plus AWS public cloud capacity.