Shipbuilder Will Triple Infrastructure Capacity, Reduce TCO by 50 Percent, with Virtualization
Harland and Wolff, an Irish shipbuilder that has constructed 1,700 vessels over its 150-year history, reorganized in the face of tough economic times, emphasizing engineering services and shorter-term projects. That called for a more agile and available IT infrastructure to support new business needs. The company responded by switching its primarily physical environment to a virtual one, using Microsoft Hyper-V technology. Agility is indeed up—workload testing, for example, has gone from days to minutes—and availability is 99.999 percent. A business continuity plan, affordable for the first time, has been added. When the migration is complete, total cost of ownership (TCO) will be reduced by 50 percent, saving more than €100,000 (U.S.$129,000). The capacity of the infrastructure will triple from 14 physical servers to 45 virtual machines, with no increase in IT personnel, for a productivity gain of 200 percent.
The company has continued to demonstrate its emphasis on innovation—not only in ship design and manufacture, but also in an increasing range of related pursuits. Harland and Wolff has distinguished itself in everything from the manufacture of tanks, aircraft, and antiaircraft guns to locomotives, offshore drilling rigs, semisubmersibles, and carrier vessels. In 2002, facing major changes in the economics of its markets, Harland and Wolff turned its penchant for innovation on itself—restructuring itself into a company focused on ship design and engineering, ship repair and conversions, oil rig construction, and the manufacture of offshore wind farms for renewable energy.
Changing Business, Changing TechnologyWith the restructuring of its business, Harland and Wolff found its technology needs changing as well. Instead of focusing on years-long construction projects, the company was now more often handling projects with weeks-long schedules. The traditional eight-hour workday schedule was increasingly replaced by work that might need to be done at any hour. These changes required technology systems far more agile and available than the aging VMS-based midrange computers on which the company depended.
A Striking Population IncreaseHarland and Wolff responded by moving from the VMS environment to a more modern client/server architecture including seven Dell server computers running the Windows operating system, and an EMC storage area network (SAN). Over the next few years, the company’s growth led to a doubling of the server population, to 14, and the need for additional computers was apparent. An increasing server population not only meant the purchase of more machines, but also more space to store them, more energy costs to run them and air-condition them, and potential additions to the four-person IT staff to maintain them.
It also meant that a business continuity plan was more important than ever. The company had suffered a couple of outages of the system over time—nothing major, but they left the engineering staff unable to work until repairs were made. The lost productivity was a problem, as was the need to incur overtime pay to make up for that loss. The risk of missing imminent client deadlines was a greater problem.
“In this economic climate, you don’t get second chances,” says Robert McIlwaine, Group IT Executive, Harland and Wolff. “It’s difficult to win work, margins are lower, and there are fewer opportunities to expand. We needed our technology infrastructure to work for us, not against us.” At the same time, Harland and Wolff wanted to avoid unnecessary expense.
The solution promised to reduce costs and support continued growth by running more workloads on fewer physical servers. The lower cost would make it practical for Harland and Wolff to create a second environment for business continuity, and the agility inherent in virtualization technology would enable failover to that second environment with minimal disruption to users.
Not Keen on VMwareHowever, Harland and Wolff was not keen to adopt VMware technology to realize the virtualization scenario. Since moving from the VMS environment, the company had built its server infrastructure on Microsoft software, and McIlwaine and his colleagues didn’t want to invest in another technology and the staff training required for it. Nor did they want to pay for the licenses that would have been required for the level of VMware that they needed.
6 Hosts, 25 Virtual MachinesThe six physical servers host some 25 virtual machines running mission-critical software including Microsoft Exchange Server, Microsoft SQL Server data management software, and Oracle Database, plus line-of-business software supporting everything from engineering to human resources. McIlwaine expects the virtual machine population to increase to 45, hosted on the same six physical servers, when the switch to the virtualization environment is complete in 2011. Already, virtually all of the company’s workloads run on this infrastructure, with minor exceptions that require their own hardware.
The virtualization solution includes two production environments running in an active/active mode for failover and business continuity, one test environment, and one archive environment. The previous infrastructure lacked the failover, test, and archive environments. “We used to run tests in our production environment,” McIlwaine says, “and just keep our fingers crossed hoping for a successful outcome.”
All of the physical servers are organized in a single cluster. That enables a failure in any server at either site to be recovered by any other server, with no more than a few minutes of downtime. For planned downtime—for example, to implement upgrades or scheduled maintenance—virtual machines can be moved between physical hosts with no downtime. The two SANs are synchronized to similarly provide failover redundancy for each other.
To implement the switch and now manage the virtual environment, Harland and Wolff uses Microsoft System Center software, including System Center Virtual Machine Manager 2008 R2 and System Center Configuration Manager 2007. The company converted the physical environment by using the physical-to-virtual tool in System Center Virtual Machine Manager. With the virtual environment now in use, Harland and Wolff can move virtual machines from one physical server to another, again by using System Center Virtual Machine Manager. The company uses System Center Configuration Manager to implement software updates across the environment.
The initial deployment took 17 business days over the course of a month, and was completed in July 2010.
Reduces Testing from Days to Minutes; Achieves 99.999 Percent UptimeThe company gains greater agility with its virtualization solution. For example, testing new workloads on physical servers used to be a weeklong process. Now Harland and Wolff creates a new virtual machine in its test environment in 10 minutes or less, and then completes the test and removes the virtual machine in about 35 minutes. Because virtual machines are just as quick and easy to create and remove in the production environment, Harland and Wolff can expand or contract that environment as needs grow or contract. “I’ve been in IT for 35 years,” says McIlwaine, “but I’ve never seen anything operate as smoothly as the Hyper-V environment.”
The company’s ability to move virtual machines from one physical server to another means that a server can be taken offline for updates and maintenance without any disruption to users. That contributes to the high availability that Harland and Wolff needs to support more dynamic, shorter-term projects. McIlwaine says that the virtual environment has 99.999 percent uptime.
“Our engineers can do their best for our customers because our technology is doing its best for them,” he says. “We’re no longer resource-constrained. Technology is no longer an impediment to moving forward.”
Protects Crucial Assets with Business ContinuityThe same features that promote agility and availability also form the foundation for the business continuity that Harland and Wolff now has with Hyper-V. “We can lose a virtual machine without losing the cluster,” says McIlwaine. “We can lose a building and half of our environment, and the virtual machines will start up on the other half within minutes.”
The dual SAN structure means that data will also be available to Harland and Wolff staff in one facility should it inadvertently go offline in the other. The lower costs associated with the new infrastructure make it possible for the company to consider investing some of its savings in a third facility, which would provide additional business continuity in the form of low-cost disk backup. “We can’t be too careful in protecting crucial assets,” says McIlwaine. “Now we can afford to take every appropriate step to safeguard our data—and our future.”
Will Reduce Total Cost of Ownership by 50 Percent, Saving €100,000In a cost-conscious climate, Harland and Wolff needed to plan for growth and business continuity without incurring excessive cost. The company achieved this goal through its adoption of Hyper-V virtualization technology. It will expand to its planned 45 virtual machines for just 50 percent of the cost of building and maintaining a comparable environment of 45 physical servers, for a savings of more than €100,000 (U.S.$129,000).
McIlwaine estimates that the virtual environment saves Harland and Wolff about 87 percent on hardware and 25 percent on software licensing, compared with an all-physical environment. Because the time required for system management is vastly reduced thanks to fewer physical servers and the use of System Center management tools, the four-person IT staff that managed the 14-server environment will be able to manage the 45 virtual machines without any staff additions. That’s a 200 percent gain in productivity.
“Adopting Hyper-V virtualization technology means much more than saving money as we grow,” says McIlwaine. “It’s the difference between growing and not growing. We couldn’t have done this otherwise. Going virtual means we can go—and grow—into new markets.”
Microsoft VirtualizationMicrosoft virtualization is an end-to-end strategy that can profoundly affect nearly every aspect of the IT infrastructure management lifecycle. It can drive greater efficiencies, flexibility, and cost effectiveness throughout your organization. From accelerating application deployments; to ensuring systems, applications, and data are always available; to taking the hassle out of rebuilding and shutting down servers and desktops for testing and development; to reducing risk, slashing costs, and improving the agility of your entire environment—virtualization has the power to transform your infrastructure, from the data center to the desktop.