The Ireland-based Coal Marketing Company (CMC) has ripped out ageing IBM/NetApp arrays and deployed Tintri hybrid flash storage to support the sale and shipment of 35 million tonnes of coal per annum worth more than €2bn.
Tintri fought off competition from storage array suppliers including EMC, X-IO, Nimble Storage, Tegile, Dell and IBM on the basis of its ease of use and deep integration with VMware.
CMC is the commercial arm of a Colombian coal mine that ships to customers in the Americas, Europe and Asia. The company has 10,000 employees at its Colombian mine complex at Cerrejon (pictured) while the commercial arm of 35 people in Dublin and Atlanta, US, supports sales staff.
The company’s IT environment runs on VMware with 61 virtual machines on hosts split between Dublin and Atlanta. Sales staff access apps via virtual desktops on VMware View while VMware Site Recovery Manager is used to provide disaster recovery between the sites.
Storage feeling the strain
Despite the small size of CMC’s commercial arm, IT is business critical. CMC uses specialist mining enterprise resource planning (ERP) software, MineMarket from MinCom, to sell 35 million tonnes of coal to precise volumes and delivery timeframes.
It was running its systems on storage from an IBM N3300 – a rebadged NetApp FAS2020 – but that array had been in place for five years and was feeling the strain. So, IT manager Geoff Grice and his team began to look for replacement storage to support the virtualised environment.
More on hybrid flash storage
- Definition: Hybrid flash array
- Hybrid flash array: What it is and how to use it
- Big storage turns the tide in the hybrid flash array market
- Telematics firm builds datacentres on Compellent hybrid flash
- University boosts VDI IOPS by 70% with Nimble hybrid flash array
- Health trust adopts Tintri hybrid flash storage in VDI rollout
“As we virtualised more and more and added the ERP system, the system overhead just became too much for the IBM array, in terms of maintenance and support as well as reaching capacity. All of which meant latency was too much for the VDI environment,” he said.
It became clear the IBM/NetApp array was not sufficient to take the company to the next level, especially as it required intensive management attention to keep it running.
“It increasingly needed hands-on attention, to repurpose LUNs and volumes, to carve out more storage, and we just didn’t have the staff to keep on top of this,” said Grice.
His team evaluated storage arrays from HP, EMC, X-IO, Nimble Storage, Tegile, Tintri, Dell and IBM in what the IT manager described as a “virtual beauty parade” that employed remote sessions via TeamViewer to examine the configuration, management, trouble-shooting and reporting capabilities of the various products.
“Some were going in the right direction,” said Grice, “while others seemed to just throw some SSD into existing arrays.”
Ultimately, Tintri won out on grounds of performance, cost, features and support. CMC deployed three Tintri T540 arrays with 12TB of capacity each, two in Dublin and one in Atlanta, and using Tintri ReplicateVM to copy data between the sites.
Tintri captures the big picture
Tintri is one of a new breed of startups that marries low-latency flash storage with bulk capacity on spinning disk. The company came out of stealth in 2011 and builds devices that specifically target VMware users. It uses data deduplication and tiering with MLC flash and Sata drives to ensure that the vast bulk of I/O hits solid state storage.
The dual-controller Tintri VMstore arrays provide iSCSI (1Gbps Ethernet and 10GbE ports) connectivity in a 3U form factor.
Tintri specifically targets virtual machine environments. To do this it does away with volumes, LUNs and RAID groups, and maps I/O requests directly to the virtual disk. This tight VM integration lets VMstore control I/O performance for each virtual disk.
Grice said benefits include vastly reduced VDI boot times, decreased time to run SQL reports, by up to 75%, as well as much improved manageability.
“Tintri integrates with VMware at a very deep level, meaning it does not just report on disk activity, but looks at the whole picture end-to-end, including host, network and storage, and can drill down to individual VM [virtual machine] level. No other system I saw could do that as they were all stuck in the mindset of aggregates, volumes and LUNs,” said Grice.