Ragic makes overnight switch to Google Compute Engine to eliminate DDoS service disruption
Founded in 2008 and headquartered in Taiwan, Ragic, Inc. develops and sells software that enables businesses to build web database applications. Partners of these businesses may also use Ragic to develop or customise database applications for them.
The software provides online database management, collaboration, report generation, query building and file storage integration capabilities. According to Jeff Kuo, Founder, Ragic, Inc., Ragic software presents an easy to use, flexible option for businesses that want an alternative to packaged software systems that may not meet all their needs. “Vendors in the enterprise resource planning or customer relationship management system markets typically market large software packages designed to meet the needs of all the customers they want to do business with,” he says. “However, these packages can require considerable time and resources to maintain, while not actually meeting all the specific needs of a business without additional customisation work.”
According to Kuo, Ragic can easily be adapted to customers’ individual requirements and is simple enough to enables team members outside IT departments to create web-based database systems. The company’s message is clearly resonating in international markets: Ragic customers include globally recognised United States universities and technology companies.
Ragic, Inc. started marketing Ragic in 2010 and elected to do so from a multi-tenant hosted platform. However, as the platform’s profile increased, so did the volume of distributed denial of service (DDoS) attacks by individuals or groups. By late 2015, Ragic, Inc. started to experience some network instability that affected the service provided to customers.
On Christmas Day that year, disruption caused by the DDoS attacks escalated and by 2 January 2016, customers and the Ragic, Inc. support team reported they were unable to connect to the company’s servers. By the afternoon, the platform provider had effectively shut down traffic to its primary datacentre, which meant none of the Ragic, Inc. servers were visible to the company or its customers. “In the end, there was nothing we could do but wait until the provider resolved the issue,” says Kuo.
While waiting for Ragic to be restored, Kuo decided Ragic, Inc. should at least understand the implications of moving its servers to another environment. Having previously used a cloud service offered by a multinational company, he elected to quickly review the Google Cloud compute product.
“The interface for Compute Engine within Google Cloud Platform (GCP) was considerably more modern and intuitive than the interface of the equivalent product on the other service,” says Kuo. “In fact the entire cloud platform was more modern and clearly designed with much less ‘baggage’ than the alternative.”
Furthermore, Google operated a data centre in Taiwan where nearly one third (30 percent) of Ragic, Inc.’s customers are based. “This means the connections for those customers in particularly will be a lot better and for a variety of reasons, including data sovereignty and support, they prefer to have a datacentre in their own country,” says Kuo.
Finally, Ragic, Inc. welcomed the ease of accessing Compute Engine functionality that encrypted data at rest, which the company viewed as extremely important in protecting the enterprise-grade data used on its service. Compute Engine’s ease of use enabled Ragic, Inc. to migrate from its existing platform to a stable and secure environment in just one night. “We found it extremely easy to set up virtual compute instances in Compute Engine,” says Kuo. “On the night of 2 January 2016, I was able to use the brief moments that our connection to our incumbent provider was available to set up all of our servers in Compute Engine and move our data across to Google.”
Ragic, Inc. is now operating a hybrid architecture based on Google Compute Engine and incorporating storage, mail, content delivery and domain name system web services from different providers. Because the business accesses most of these services through application programming interfaces, using multiple providers has not proved to be any impediment to operating efficiently, flexibly and cost-effectively. In May 2017, Ragic Inc. was running five Compute Engine standard virtual machine instances across Taiwan, the United States and Europe to meet the compute requirements of the Ragic software.
Greater stability to improve performance and availability
The migration to Google Compute Engine has delivered a range of benefits to Ragic, Inc., including stabilising the service to improve performance and availability and blocking the threat posed by DDoS attacks. “We have had no service disruptions or degradation due to DDoS attacks – either to our underlying platform or the application itself – since we moved to Compute Engine,” Kuo points out. “We do have alert tools that notify us when we are under attack and we have not received any of these since we moved to Google about 18 months ago.
“We have been continuously improving our availability and for the most recent month, we’ve reached nearly 100 percent.”
This robust performance is also due to the fact that Ragic Inc. can add extra disk space to its virtual machine instances without scheduling downtime to complete the required work.
Faster response times
Running virtual machine instances close to large contingents of users in Taiwan, the United States and Europe has enabled users to connect to a local server to ensure faster response times. “The machines themselves are quite a bit quicker than those we were using previously, and we’ve improved response times from countries accessing our instances in the United States from up to 400 milliseconds to about 200 milliseconds,” Kuo says.
By deploying instances in Europe, Ragic, Inc. has been able to ensure that its European customers can comply with data sovereignty and governance requirements. Ragic, Inc. is extremely pleased with the stability of its architecture and particularly Google Compute Engine since the start of 2016. The business is evaluating whether other Google Cloud services such as Google Natural Language API can deliver further efficiencies to its operations.
“We don’t really have many issues that Google should address and it is extremely quick at providing status updates and information as we need them,” says Kuo.