About The GPT Group
The GPT Group is one of Australia’s largest diversified property groups with $15.2 billion in assets. The company, headquartered in Sydney, Australia, has 500 employees and has been listed on the Australian Securities Exchange (ASX) since 1971. GPT manages two wholesale funds and a real estate portfolio that includes retail, office, logistics, and business park assets. In 2013, GPT was ranked in the top one percent of global real estate companies on the Dow Jones Sustainability Index (DJSI).
To create a foundation for growth, GPT wanted to lower costs and extract as much value as possible from its internal resources, especially in the IT department. At the time, GPT was using two hosting providers in Sydney to run its core business applications and databases. The production data center hosted 10 racks of equipment and fiber links that provided fast but expensive storage area network (SAN) replication to a secondary data center running four racks of equipment.
The IT department was tasked to reduce costs by 20 percent, deriving a large share from infrastructure savings. Additionally, IT had to ensure that its systems — including an accounting system based on an SAP ECC enterprise resource planning package — maintained the same level of performance as its existing hosted data center environment.
“This target effectively brought plans we had to rationalize our infrastructure forward by 12 to 18 months,” says Michael Lockhart, IT Infrastructure Manager. “We also had some major data center contracts at our hosting centers due for renewal in 2013, so it was an ideal time to review our approach.”
When extensive analysis and modeling revealed that the firm could achieve considerable cost reductions by using a cloud service provider, GPT began to consider migrating to the cloud. Security reviews by external consultants and a wide-ranging self-assessment indicated that there were cloud providers with more rigorous security than GPT’s existing hosted data center environment. GPT began to evaluate cloud service providers, and given the scope of its proposed migration, wanted to be confident that its chosen cloud provider would be around for the long term and have the cloud as its core competency. With these requirements, GPT identified a short list of three cloud service providers, including Amazon Web Services (AWS).
Why Amazon Web Services
After rigorously testing and analyzing different cloud environments, the firm decided to move its infrastructure to AWS because of low cost, scalability, and features such as the fast deployment of new instances.
GPT was also extremely impressed with the AWS culture and record of continuously adding new features and functions. “We were looking to the future as well as our existing requirements,” says Lockhart. “AWS gave us confidence that as our needs change and grow, the company will have the products available to support us.”
The IT infrastructure team presented the business case for moving to the AWS Cloud to its senior leadership team in late 2012 and received approval to proceed. GPT brought in a project manager with data center experience and a solutions architect with previous cloud design experience to work with AWS to develop the high-level architecture needed to meet GPT’s requirements for the deployment. “This helped us reduce the time required to come up with the design and the complexity of the architecture itself,” says Lockhart.
GPT lifts and shifts its data center to the AWS Cloud
GPT migrated almost its entire data center to the AWS Cloud. “About 70 percent of the migration was a straight lift and shift into the cloud,” says Lockhart. The remaining 30 percent involved rebuilding systems, including SAP application servers and Active Directory (AD). Lockhart comments, “We rebuilt our SAP system and moved it to AWS in one weekend and moved Active Directory across in one day.”
The AWS Cloud runs nearly all core GPT systems, including its reporting environment and authentication system, SAP ECC, Active Directory, SQL databases, and Microsoft SharePoint 2007. AWS Direct Connect provides a direct link from the AWS Cloud to the GPT’s Wide Area Network (WAN). Separate interfaces and middleware between the SAP system and the property management system enable GPT to complete critical tasks such as running rent rolls and invoicing. The only system that GPT didn’t move to the cloud was its property management system. “We still run this system on an AS/400,” says Lockhart. “There’s no technology available that allows us to virtualize this system at the moment, so it wasn’t possible to undertake the migration.”
GPT runs on 112 Amazon Elastic Compute Cloud (Amazon EC2) instances. Amazon Elastic Block Store (Amazon EBS) provides block-level storage for over 300 disk volumes comprising 37.5 terabytes (TB) and balances traffic across the firm’s three Microsoft SharePoint instances. The environment resides within an Amazon Virtual Private Cloud (Amazon VPC) running across AWS Availability Zones in the Asia Pacific (Sydney) Region, and includes a demilitarized zone for Internet-facing reverse proxy and web servers. Figure 1 demonstrates GPT’s environment on AWS.
Figure 1. The GPT Group on AWS
By February 2014, GPT was using over 2,300 Amazon EBS snapshots to provide point-in-time backup for the Amazon EBS volumes, while managing 7.2 TB of inbound data and 33 TB of outbound data. GPT uses Amazon Support, Business level to support its environment and Lockhart says, “Because our AWS infrastructure is very stable, most of our support calls are designed to tune the environment to get more from our systems. We’ve found AWS Support to be extremely responsive and perform the way a support operation should.”
GPT estimates a wide range of benefits moving to the AWS Cloud, including lower costs, better security, consistent application performance, and the ability to respond faster to growth opportunities.
“While we’re still investing capital in acquiring Amazon EC2 Reserved Instances that make up about 80 percent of our overall environment, we still expect to save up to 60 percent on an hour-by-hour basis,” says Lockhart. “The firm is tracking operating expenditure savings and we are about 25 percent ahead of budget at this stage.” Including capital expenditure cost avoidances — servers, storage area networks and other equipment due for renewal in 2013 — the company expects to recoup its outlay on the project in less than 12 months.
Prior to moving to AWS, GPT conducted performance testing for its SAP and SharePoint implementations. “One of the project’s critical success factors was ensuring we would maintain at least the same level of performance as our previous hosted data center environment,” says Lockhart. In each case, the firm found the time taken to complete those tasks in AWS matched the time taken in its previous infrastructure. “We found, for example, that running an SAP general ledger line item report which took three minutes in our previous environment took three minutes in the AWS Cloud.”
Furthermore, GPT believes that its applications, infrastructure, and data run more securely on AWS than its previous hosted environment. “There was no way we could achieve the security certification levels that AWS has,” says Lockhart. “We have great confidence in the logical separation of customers in the AWS Cloud, particularly through Amazon VPC, which allows us to customize our virtual networking environment to meet our specific requirements.”
The move to AWS means that the IT infrastructure team is now well positioned to support GPT’s growth strategy. “We no longer have to wait six to eight weeks for hardware to turn up,” says Lockhart. “If there was merger and acquisition activity involving GPT, we know that we could respond by building new infrastructure within hours.”
Now that the migration phase of the project is complete, GPT is poised to optimize its use of the AWS Cloud. The company plans to use AWS scripts and APIs to automate functions, which will allow the IT infrastructure team to focus on projects and activities that generate value for the business.
“With AWS, the business is extremely confident that the IT function can meet its current and future needs,” says Lockhart. “We’re looking at ways we can better use the cloud environment, including services such as Amazon Relational Database Service (Amazon RDS). With the new services AWS is delivering, we’ll identify more opportunities to operate more efficiently and effectively.”