Innovaccer Lowers Costs by 33% and Reduces Management Overhead by 65% by Modernizing on AWS

Innovaccer's mission is to connect and curate the world’s healthcare information, enhancing accessibility and usability. The enterprise offers software as a service (SaaS) solutions that aggregate data across systems and healthcare settings to deliver standardized, unified patient records. Its platforms are advancing the transition from fee-based to value-based care services, prioritizing the patient experience and reducing healthcare costs over time. This shift represents a significant opportunity to enhance patient outcomes and efficiency in the healthcare industry.

Currently, Innovaccer supports around 96,000 providers across 1,600 locations, predominantly in the United States. Its artificial intelligence (AI)-powered platform runs complex analytics on hundreds of terabytes of data each day. Born on Amazon Web Services (AWS), the company uses a variety of database and data processing technologies to manage large volumes of data efficiently.

In response to the challenges posed by a growing volume of patient records in varied formats, Innovaccer set out to modernize its technology infrastructure. This strategic update was aimed at enhancing system efficiency and managing costs more effectively, ensuring that the company continues to meet the increasing demands of its enterprise-level customers, including recent additions like Kaiser Permanente.

Innovaccer began modernizing its underlying database architecture to enhance throughput as it onboarded enterprise-level customers. The company uses SQL scripts to extract, transform, and load (ETL) processes across its custom data pipelines, with numerous jobs requiring complex SQL joins. Data feeds originate from 40‒50 sources in both structured and unstructured formats, including electronic medical records and medical device data. 

Additionally, Innovaccer adopted Amazon Aurora as a PostgreSQL-compatible relational database to manage online transaction processing (OLTP) and complex online analytical processing (OLAP) workloads. Arun Buduri, vice president at Innovaccer, says, “The combination of OLAP and OLTP architecture on AWS has effectively kept our data processing costs in check. This system is much more efficient and reliable compared to our previous database architectures.” Innovaccer also improved performance and reduced costs by migrating all of its Amazon Relational Database Service (Amazon RDS) instances to AWS Graviton Processors.

Throughout the modernization project, Innovaccer met weekly with its AWS account team and solutions architects to explore how AWS services could enhance its business. One matter that emerged from these meetings was Innovaccer’s high input/output operations per second (IOPS) usage, stemming from intensive transactional reads and writes on its databases. When AWS introduced Amazon Aurora I/O Optimized for IOPS-intensive applications, the account team proposed Innovaccer adopt the service to further optimize costs. Since adopting the service, the company has saved 45 percent of its monthly Amazon RDS cost.

Innovaccer also implemented Amazon Elastic Kubernetes Services (Amazon EKS) and Amazon Elastic Container Service (Amazon ECS) to manage its newly containerized workloads. The AWS account team organized training sessions on these technologies for Innovaccer’s site reliability engineering and products team, benefiting over 300 employees. These employees have given overwhelmingly positive feedback, noting that the training has significantly simplified their work.

In addition, Innovaccer has invested in managed services like Amazon OpenSearch Service as part of its modernization efforts. Bill Burton, vice president of cloud engineering at Innovaccer, says, “We've achieved a 65 percent reduction in management overhead, helping our team to concentrate on high-impact projects rather than manually managing infrastructure. We've also noticed improved reliability with AWS managed services compared to our previous self-managed approach.”

After modernizing its database, analytics, and compute architecture over a period of three years, Innovaccer has reduced overall cloud overhead costs by about 33 percent. The business has also improved performance by 30 percent, while continuing to grow its business year-over-year. Innovaccer has been able to pass these savings onto its customers, keeping costs low to align with its mission of advancing value-based care.

Thanks to the use of more managed services on AWS, Innovaccer’s team can now dedicate more time to new projects. Burton explains, “With production automation and improved visibility, our teams are able to concentrate on important projects that previously might not have been feasible due to time constraints.”

A focal point of Innovaccer’s production agenda is AI. The company is currently hosting vision models using NVIDIA Triton Inference Server in Amazon SageMaker and generative AI workflows on Amazon Bedrock to integrate AI into all its products. “The AI and ML tools available on AWS are remarkable for simplifying model deployment and auto-scaling,” explains Buduri. “This significantly reduces the burden on our ML team, as they no longer need to build additional services on top of the models.”

Nilav Baran Ghosh, director of Data Science at Innovaccer, adds, “Amazon Bedrock gave us a No Code/Low Code approach to quickly establish an RAG [Retrieval-Augmented Generation]−based system with leading foundation models for searching through our multimodal content. This approach helped us to get up and running in just a few minutes, significantly accelerating our build process without any upfront cost commitment."

Reflecting on Innovaccer’s ongoing modernization and development, Burton concludes, “AWS has been deeply engaged with us, demonstrating a commitment to our success that matches our own.”