Customer Overview
Founded in 2007, Labguru, the flagship product of BioData, is a versatile and unified cloud-based data management platform that supports a wide range of laboratory and research activities. Trusted by over 750 global customers – including national research institutes, academic institutions, pharmaceutical companies, and innovative biotech startups – Labguru plays a pivotal role in modernizing lab and research management. The platform integrates an Electronic Lab Notebook (ELN), Laboratory Information Management System (LIMS), inventory management, automation tools, and an advanced informatics platform into one easy-to-use solution.
By centralizing data within secure, cloud-based software, Labguru streamlines lab operations, prevents data loss, and accelerates research outcomes, ultimately saving both time and money. The platform is accessible from both desktop and mobile devices, ensuring that researchers can manage their work efficiently, regardless of their location.
Labguru’s broad application in research and laboratory management, biotechnology, pharmaceuticals, and related fields makes it an indispensable tool in the life sciences sector.
Challenge
Labguru relies on AWS Elastic File System (EFS) to manage mission-critical storage that underpins its application and serves a diverse global customer base. Initially, Labguru deployed EFS using the default bursting throughput mode, which provided sufficient performance for their workloads. AWS EFS, known for its scalability and seamless integration with other AWS services, was well-suited to Labguru’s need for reliable, cloud-native storage that could grow in tandem with their business.
However, as Labguru’s usage patterns evolved, they began experiencing performance bottlenecks with the burstable throughput mode, particularly during periods of high demand. To address these issues, the engineering team migrated to EFS Elastic Throughput mode, which is optimized for spiky and unpredictable workloads. While this transition resolved the performance issues, it also led to a significant and unexpected increase in costs – 202% month-over-month (MoM) following the migration, escalating to 305% two months later, despite the storage volumes remaining relatively constant.
The unexpected cost increase placed pressure on Labguru’s operational budget, necessitating a thorough investigation into the root cause of the issue. Although Labguru’s operations continued without disruption, the increased costs and the resulting financial analysis diverted the engineering team’s attention from critical development projects. This misalignment of focus hindered their ability to drive innovation and meet strategic goals.
Proposed Solution & Architecture
In response to the rising costs associated with the Elastic Throughput model on AWS EFS, Labguru sought to optimize its AWS expenditures without compromising performance. To achieve this, they engaged with GlobalDots, a leading global cloud solutions integrator and FinOps service provider. The GlobalDots team conducted an in-depth analysis using AWS Cost Explorer’s advanced filtering capabilities to track the cost spikes and correlate them with EFS usage patterns. They also leveraged AWS CloudWatch to monitor filesystem performance metrics, identifying that while the Elastic Throughput model efficiently handled spiky workloads, the maximum throughput demands were capped at a certain threshold, suggesting that an alternative approach might yield better cost efficiency.
After analyzing the performance data and cost trends, GlobalDots proposed a shift from the Elastic Throughput model to the Provisioned Throughput model for certain Labguru’s EFS workloads. This model enables customers to specify the throughput capacity they require, allowing them to tailor performance to their actual workload demands.
The team conducted a detailed comparison between the Elastic and Provisioned Throughput models, taking into account Labguru’s specific requirements, including storage capacity, read/write operations, and infrequent access storage tier patterns. They recommended provisioning a throughput capacity that included a 20% headroom above Labguru’s maximum observed workload throughput (measured in MB/s), ensuring that performance requirements would be met even during unexpected spikes.
By inputting these parameters into the AWS Pricing Calculator, the team determined that switching to Provisioned Throughput would significantly reduce costs – potentially offering a ~50% reduction compared to the Elastic Throughput model. Although this model wouldn’t achieve the lowest costs experienced under the original burstable model, it represented an optimal balance between cost and performance for Labguru’s needs.
To ensure smooth implementation and ongoing cost management, GlobalDots recommended setting up CloudWatch dashboards and alerts to monitor filesystem performance and detect any potential issues early. Additionally, they suggested using daily-granularity reports in AWS Cost Explorer to track the financial impact post-migration and adjust the provisioning as necessary.
The solution primarily involved the following AWS services:
Amazon Elastic File System (EFS): Initially configured with Elastic Throughput, the focus was on identifying a more cost-effective model while maintaining necessary performance levels.
AWS Cost Explorer: Used extensively to analyze cost data, identify trends, and forecast potential savings.
AWS CloudWatch: Deployed to monitor real-time performance metrics, helping to assess the suitability of various EFS throughput models.
AWS Pricing Calculator: Utilized to model different EFS configurations, providing a detailed cost analysis of the proposed changes.
Results:
Labguru achieved a 60% monthly reduction in AWS EFS costs with no performance degradation. The engineering team avoided disruptive architectural changes, allowing them to focus on core projects. The solution reaffirmed Labguru confidence in their AWS architecture while significantly improving cost efficiency.
Lessons Learned:
This case underscores the importance of avoiding a one-size-fits-all approach in cloud resource management. While Elastic Throughput suits spiky workloads, benchmarking against alternative models like Provisioned Throughput can yield substantial cost savings. Communicating potential impacts to engineering and development teams is a FinOps best practice that ensures alignment and optimized outcomes.
