Improving Performance by 30% for 400 Million Daily Transactions Using Amazon EBS with McAfee

McAfee was founded in 1987, and the company is focused on protecting people, not just devices, empowering them to live securely through integrated, intuitive solutions. McAfee secures its customers’ digital lives by protecting identities, providing virtual private network connections, and enhancing privacy. When McAfee saw rapid growth in 2020, the company had already been exploring cloud services as it planned for future expansion, and it was using AWS services for some of its new products.

After a positive initial experience, McAfee decided to migrate its business-critical SQL database to AWS. “We chose AWS because our team already had the necessary skill set, the costs were lower than other offerings, and AWS has multiple AWS Regions for our operations,” says Srinivas. After several months of planning, McAfee migrated the database to Amazon EBS io2 Block Express volumes, which offer block storage in the cloud with four times higher throughput, input/output operations per second, and capacity than io1 volumes, along with sub-millisecond latency. McAfee is seeing significantly better speed and performance since the migration. “Using our on-premises infrastructure, reporting took 4 hours, but we reduced that to as little as 45 minutes by migrating to AWS,” says Lohith KS Chalam, lead database administrator at McAfee.

Throughout its migration, McAfee worked closely with AWS to verify that it had the right plan in place. The company engaged in multiple reviews using AWS Well-Architected, which cloud architects can use to build secure, high-performing, resilient, and efficient infrastructure for a variety of applications and workloads. It also bounced ideas off of AWS support professionals to find the optimal services for its use case. McAfee used Amazon Elastic Compute Cloud (Amazon EC2), which offers secure and resizable compute capacity for virtually any workload, to power its critical workloads in the cloud. More specifically, to optimize performance, it used the AWS Nitro System, a combination of dedicated hardware and lightweight hypervisor for faster innovation and enhanced security. “We shared our requirements with AWS for our critical workload and data warehousing workload, and AWS recommended Nitro-based Amazon EC2 instances, which have met our needs,” says Saravanakumar Karunakaran, senior architect at McAfee.

Using Nitro-powered Amazon EC2 R5b Instances, which are the next generation of memory-optimized instances for Amazon EC2, as well as Amazon EC2 X2i Instances, which offer up to 50 percent better performance than X1 instances for memory-intensive workloads, McAfee boosted performance by 30 percent across 400 million daily transactions. “After migrating, our applications were flying along, and we haven’t seen any issues with storage or computing,” says Saravanakumar. Although McAfee enhanced performance across many dimensions, latency improvements in querying have been particularly impactful for both internal teams and customers. For example, login times have reduced significantly, leading to increased customer satisfaction. Also, overall query times have fallen by about 20 percent, grounding a more seamless experience across McAfee’s global customer base.

Now that its workloads are running on AWS, McAfee teams can more easily carry out enhancements without disrupting service for customers. “With our legacy infrastructure, failovers used to take 5 minutes,” says Srinivas. “Now, we’re consistently seeing less than 90 seconds of failover time.” Moreover, backup time for critical workloads has fallen by 50 percent using Amazon EBS io2 Block Express volumes. For noncritical workloads, McAfee uses Amazon EBS General Purpose Volumes, which are solid state drive volumes that offer balanced cost and performance for a majority of workloads, to deliver greater performance and reduce costs.

McAfee achieved significant overall cost savings by migrating to AWS. “When we look at hardware costs and add the operational costs of the teams involved in running an on-premises data center, we have saved significantly by running our systems in the AWS Cloud,” says Srinivas.