RMS Cloud Overcomes “Noisy Neighbor”, Improving Database Performance by 30% on AWS

For more than 40 years, RMS Cloud has provided an all-in-one property management platform to hospitality businesses in Australia and internationally. In fact, more than 7,000 properties in 70 countries currently use RMS software. RMS Cloud has been in a process of digital transformation, constantly looking for new ways technology can deliver cost-efficient, resilient solutions at scale. Since 2011, the company has used a multi-tenant software model on the cloud. 

As data volumes grew, the software provider faced challenges with its cloud-based database management service, often experiencing downtime during traffic surges. Complications with one customer would often impact other tenants sharing the same database instance, leading to widespread service downtime. To address potential client complaints, RMS Cloud resorted to assigning customers to individual single-tenant instances—a costly solution that undermined the benefits of a multi-tenant model. 

In January 2023—one of the busiest months for the Australian hospitality industry—the company’s cloud provider suffered a significant outage that resulted in nearly 8 hours of downtime overnight. RMS Cloud is rapidly growing in Asia and Europe, and the outage did not bode well for this expansion. More importantly, engineers had little to no support from the provider during the event. RMS Cloud soon decided to switch to a provider that could offer higher resilience and also address the “noisy neighbor” issue with its multi-tenant database model. 

RMS Cloud found the support it was seeking in Amazon Web Services (AWS). The company’s AWS account team connected it with AWS Partner Slalom to build a modern cloud strategy. What began as a straightforward migration grew to encompass higher availability, better performance, and lower database cost on AWS. 

Slalom initiated a comprehensive cost and technology analysis using the AWS Well-Architected framework, spending nearly three months designing and stress-testing new, customized architecture before migrating any production workloads. The assessment was fully funded by AWS as part of the AWS Migration Acceleration Program (MAP), which ensured technical acceleration and commercial benefits throughout the migration.

Andrew Buttigieg, technical services manager at RMS Cloud, says, “The transition was very smooth. Slalom and AWS have both been very responsive throughout the process, with response times from a few minutes to an hour at most.” 

As part of the AWS Enterprise Support plan, RMS Cloud benefits from AWS Infrastructure Event Management (IEM) during migration waves. By February 2024, approximately six months after the first wave, RMS Cloud will have migrated all 4,300 databases, totaling over 2TB of data, to Amazon RDS for SQL Server. RMS Cloud also seamlessly moved its Kubernetes workloads to Amazon Elastic Kubernetes Service (Amazon EKS) and is using Amazon Elastic File System (Amazon EFS) for serverless storage.

Along with the migration, RMS Cloud has implemented additional security features such as Amazon Inspector to detect software vulnerabilities and centralize security alerts. Data is encrypted both in transit and at rest, and backups occur automatically within AWS Regions in Australia. “We can avoid security drift thanks to the thorough architecture review we’ve done prior to migration,” adds Buttigieg. “With the modernization of our database, we’ve built in additional encryption for PII [personally identifiable information].”

By migrating to AWS, RMS Cloud achieved a performance gain of 20–30 percent for its databases. Amazon RDS offers a lower CPU utilization rate compared to RMS Cloud’s previous provider, which means the company can run smaller machine sizes for a similar customer pool and save on cost per customer. 

The “noisy neighbor” issue that was causing downtime has also been resolved. Despite facing similar spikes in demand, these are now swiftly managed on Amazon RDS, preventing disruptions from spreading to other customers on the same server. Today, RMS Cloud observes a maximum delay of 15–20 seconds, in contrast to the 5- to 6-minute delays experienced with the previous cloud database. “We’re not babysitting the SQL servers anymore, which helps us really focus on our application development rather than database management,” Buttigieg adds. 

Additionally, RMS Cloud uses service-level granular reporting via AWS Cost Explorer, empowering the business to visualize, comprehend, and manage AWS costs and usage effectively. This meticulous monitoring of costs and consumption has facilitated a budget-controlled migration and established a solid foundation for future modernization initiatives.

Most importantly, RMS Cloud feels confident in the support they receive from AWS. Peter Buttigieg, managing director at RMS Cloud, says, “Our business requires 24/7 uptime. While challenges are inevitable, the ability to quickly connect with experts who understand our needs and respond rapidly has been a game changer for us.” Next on its cloud roadmap, RMS Cloud plans to implement Amazon Aurora for PostgreSQL as a fully managed database service.