With a storage environment ripe for replacement and nearing the end of its lease, the CIO of Aussie earth moving giant, Komatsu, took the opportunity to step back and review the organisation’s IT infrastructure.
In-depth: Data centre migration guide.
The company supplies construction, mining and heavy equipment across Australia, New Zealand and New Caledonia with 25 branches, 35 service depots and employs about 1300 staff.
Addressing attendees at the EMC Inform 2011 Conference in Sydney, Ian Harvison said the company in 2009 found it had the opportunity change up its IT infrastructure
Previously the company had a large SAP footprint which was primarily based on an itanium service from Dell, the company’s infrastructure partner, which in time had become both difficult and costly to replace as Dell had ceased making them.
Komatsu’s Fairfield-based data centre also needed work, requiring a new Uninterrupted Power Supply (UPS) and a new air conditioning system to continue to serve the company's requirements.
Harvison noted the company was no stranger to virtualization, already enlisting EMC’s VSphere 3 which, throughout 2008 and 2009, had reduced the physical server count from 90 to just 28.
Once the decision was made to find an experienced provider that could offer an Infrastructure-as-a-Service (IaaS) private Cloud model without still being in the experimental phase, Telstra and Fujitsu were shortlisted.
The company deployed Telstra’s IaaS service in a $35 million deal as it already ran on an SAP environment in the Cloud space which allowed Harvison to observe how it would work for Komatsu.
“To me this is different from a normal outsourcing model because we’re not actually letting go of the ownership or the management of our core applications,” Harvison said.
“As the business has demand for additional capacity either server or storage, we can go to Telstra and have that turned on within hours.
“Previously it would actually take us six to eight weeks to go out acquire the hardware, build it and integrate it with the business so this makes it much more flexible, much more agile."
According to Harvison, the move to Telstra’s IaaS platform has enabled the company to migrate from the Itanium SAP kernel over to an x86 kernel further virtualizing the SAP footprint with just nine physical servers remaining.
“Virtualizing itanium is challenging, it was actually one of the areas on our journey of consolidation and presented a road block for us in decreasing the number of servers that were running in our organisation.”
The company closed down its Fairfield data centre on 26 December last year following the migration of 100 per cent of its data into Telstra’s private Cloud, for which its primary data centre is located in Sydney and its disaster recovery (DR) site in Melbourne.
The ability to have all of its virtual servers mirrored into the DR site in Melbourne is one of the key benefits, Harvison said, compared with the previous environment which could only run the business for a period of time on a reduced capability.
“The data is actually mirrored between the two sites so that if we have an incident we can fail over the DR site in Melbourne without any flow on effects to the business," he said.
Harvison also detailed the project undertaken over the Easter break earlier this year, of a simulated natural disaster, following the recent Australian disasters as well as those in Japan, where Komatsu’s parent company resides.
“We took the five days and declared an incident at 10pm on the Friday night that we had lost the Pitt Street data centre, then we started with Telstra to work through migrating to our DR facilities,” Harvison said.
“The Fujitsu mainframe we actually failed over to the cold site at North Ryde and we started the process of recovering the mainframe environment, we’d never done it on the SAP environment.”
Some of the company’s main customers, like Rio Tinto, don’t care about such incidents, Harvison said, as they run 24/7 and have a contract ensuring there can always be parts delivered to their site.
“We actually ran our business from a transactional point of view on the DR environment and then we did what most companies don’t do which was file the whole our company back over to the main environment," he said.
“It was a success and the only reason we could do that was because we had the Infrastructure-as-a-Service model in the Cloud.”
Follow Chloe Herrick on Twitter: @chloe_CW
Follow CIO Australia on Twitter: @CIO_Australia