Rapid growth in our refund issuing business led to a capacity issue with our existing infrastructure. Our ability to scale rapidly in our own on-premise environment was limited by data centre space constraints. The Business Intelligence (BI) system was identified as the only non-revenue system that would free up capacity, and so it was shut down to protect the revenue systems.
The initial engineering estimates for building a new BI infrastructure were so high, that there was a real danger of losing the BI capability. I challenged the BI team to use a cloud service provider as a way of staying in business without dependency on our limited internal resources. We evaluated a number of options and chose AWS as our cloud platform.
As we were successful with moving BI to the AWS cloud, we decided to make it a permanent move, and created a Reporting Factory to handle our Merchant Reporting, management dashboards, and other scheduled reporting tasks. These tasks are classic peak computing applications - high resource utilisation for short periods of time - and lend themselves to the cloud environment. The BI Reporting Factory is running on AWS and is accessed by all the 38 locations worldwide. We are also hosting our corporate website, customer-facing websites, and our media content management systems on AWS.
Since the BI team is located in Singapore, we needed a vendor who could work with us locally, while still providing facilities in our key markets around the world. As the BI team was not engineering focused, it was also important that the platform be easy to use and configurable by non-engineers. AWS met those criteria, and we started using AWS in 2010.
The technology services on the AWS Cloud platform that we are using to build and run our BI Reporting Factory include Amazon Elastic Compute Cloud (Amazon EC2) and Amazon Simple Storage Service (Amazon S3). The main benefit was ease of use, and the ability of staff with "normal" computing skills to access and configure the setup.
No, the reason why we chose AWS was its ease of use which allowed non-infrastructure specialists to configure the platform.
The major benefits AWS brings to Global Blue are speed, capacity and cost reduction. Our initial motivation was to restore BI services in the face of a capacity crunch in our data centre. The ability to move a key system to the cloud and get back into production was important for our business. More importantly, we avoided significant costs with our BI Reporting Factory - $800,000 in CAPEX (capital expenditure) and $5,000 to 8,000 per month in OPEX (operating expenditure) - which would have otherwise gone to on-premise infrastructure expenses.
Beyond the cost savings, as mentioned above, an important consideration in choosing AWS was the ease of use and its flexibility.
Global Blue's next cloud initiative is to use the cloud to enhance our disaster recovery capability. We have been virtualised since 2007, so the cloud is an attractive option for the way we operate. We apply a test to any new system proposal that extends the old 'build versus buy' to 'build, buy, or cloud.' Our evaluation of cloud for disaster recovery is currently underway. Longer term, we are working on moving to high availability designs that take advantage of both internal computing resources and cloud services in a hybrid arrangement.
We have appreciated the willingness of AWS to work with us to explore new and innovative uses of their platform. Having a vendor's technical staff available to consult with is critical to moving projects forward.