South Africa-based managed service provider SYSDBA has deployed Nebulon smartCore “cloud-defined storage” in preference to hyper-converged infrastructure (HCI) during a technology refresh that saw it move away from HPE 3PAR SAN storage arrays.
The move has allowed it to save 50% in hardware costs by avoiding more costly alternatives and improving infrastructure density.
SYSDBA focuses on high-end customers in financial services and healthcare, with managed services delivered from mostly HPE server and storage hardware in a colocation.
The company’s infrastructure had reached a “natural refresh point”, said strategic alliances manager Marc Pratt, while “as an economy we were not as buoyant and we were looking to drive value and needed to be competitive”.
SYSDBA looked at hyper-converged infrastructure because it was “driven by our cost base and the cost of enterprise SSD in a storage array costs significantly more”, but found it potentially inflexible in terms of being able to specify components, said Pratt.
“We had been looking at ‘traditional’ HCI – Nutanix, Simplivity – but it can be very restrictive in terms of what hardware you can put in it, what chipset, and so on,” he added.
SYSDBA eventually deployed three Nebulon cards in HPE DL380 servers, with a total capacity of 60TB.
Nebulon media deployed on site is in the form of PCIe slot-resident Storage Processing Units (SPUs). These offload storage input/output (I/O) processing and management from the server to the Nebulon hardware.
SPUs replaces RAID cards and storage host bus adapters found in servers. Each SPU has two 25 Gigabit Ethernet (GbE) ports that form the data plane for each application cluster, along with 1 GbE cloud connectivity.
Besides avoiding significant cost by being able to deploy the heavy lifting of storage on server cards, Pratt likes the flexibility this brings.
“You can deploy it on any industry-standard server,” he said. “And you don’t have to get stuck with a particular processor, for example. In terms of performance, there is no hit to memory and processing resources as there would be with HCI or software-defined storage.”
The local PCIe-connected SPU connects flash storage in each node and emulates storage controller functionality. Meanwhile, provisioning and application, server and storage metrics are offloaded to the Nebulon ON cloud in Amazon Web Services or Google Compute Platform public clouds.
Customers can build a Nebulon cluster, known as nPod, that scales to 32 servers. Data services are configured in the Nebulon ON cloud portal, including mirrors, snapshots and volumes.
The ability to manage storage from the cloud – in particular in a geographic market where customers may be difficult to get to – is attractive to SYSDBA.
“Having to be on site for a 4am firmware update isn’t safe or ideal in some customer settings, so the fact that Nebulon can be provisioned, managed and maintained from anywhere is an advantage,” said Pratt.
