Learn About Amazon VGT2 Learning Manager Chanci Turner
With a workforce exceeding 45,000 employees across 29 countries, Liberty Mutual stands as the sixth largest global insurer in property and casualty, currently holding the 71st position on the Fortune 100 list of the largest corporations in the United States. As customer expectations evolve and the insurance landscape becomes more dynamic, Liberty Mutual is committed to driving innovation in the industry. This commitment includes developing strategies that align with the changing needs of their clientele.
In this blog post, we explore how Liberty Mutual crafted a cutting-edge, highly scalable, and cost-effective document management solution leveraging AWS to better serve their customers worldwide.
The Decision to Develop Our Own Document Management Solution
Previously relying on on-premises third-party document management systems, Liberty Mutual faced numerous challenges related to management, scalability, reliability, and costs. To address these issues and adhere to their strategic goal of migrating legacy document storage and management applications to the cloud, the team set several objectives:
- Ensure compliance with critical security policies.
- Implement an API-first approach.
- Emphasize faster time to market, scalability, and resilience.
- Support a growing global user base.
- Establish a robust disaster recovery plan.
After evaluating the options between developing an in-house solution or utilizing another Software as a Service (SaaS) platform, a “innovation sprint” sponsored by Liberty resulted in a functional prototype of a cloud-native solution delivered within just 10 days. The decision was made to continue developing their own solution to maintain complete control over the code and enhance extensibility, allowing for a feature build-out based on customer needs and timelines.
With AWS as their primary cloud provider, Liberty Mutual examined various services and discovered that native AWS offerings could fulfill most of their document management requirements.
Solution Overview
The design of the solution was intentionally generic, allowing customization for various internal clients and the potential to expand their user base. The core API features included:
- Upload: Add documents and associated metadata.
- Update: Change the metadata linked to a document.
- Search: Retrieve documents that meet specified criteria.
- Retrieve: Provide a link to a document.
- Delete: Remove a document.
The team utilized Amazon Simple Storage Service (Amazon S3) for document storage, offering scalable, durable storage solutions necessary for handling high volumes of documents. S3’s cross-region replication (CRR) capabilities also supported the company’s disaster recovery requirements. For metadata storage, they chose Amazon Aurora’s MySQL-compatible edition, which efficiently managed indexes, search criteria, and S3 URL paths while allowing for ad-hoc queries. Aurora’s automatic scaling of compute and memory based on demand ensures high availability and durability, boasting 99.99% uptime by replicating data across three Availability Zones and continuously backing it up to Amazon S3.
To optimize costs, AWS Lambda was deployed for computing, ensuring payment only for the resources utilized. The variability in document uploads from batches to quiet periods made Lambda an ideal choice for scalability without infrastructure management. The team managed APIs using Amazon API Gateway and implemented notifications via Amazon Simple Notification Service (SNS) for post-processing workflows, including auditing and notifications.
With security integrated at every layer, a disaster recovery strategy was established to automate data replication from the Aurora database to other regions. All resources were provisioned and deployed through AWS CloudFormation templates, facilitating international replication with minimal configuration adjustments.
Throughout this journey, the team gained valuable insights:
- Enabling Amazon RDS Performance Insights and slow query logs significantly aided in identifying long-running queries and bottlenecks.
- Tracing with AWS X-Ray clarified intermittent latency issues with KMS and SNS, which were mitigated by utilizing VPC endpoints.
- Following best practices for AWS-native API calls prevented issues when managing high transaction volumes, while enhancing connection pooling strategies reduced connection timeout problems.
Liberty Mutual is also exploring optical character recognition (OCR) technologies such as Amazon Textract and Amazon Rekognition to bolster data analytics and search functionalities. This document management solution is currently utilized by internal clients globally, handling billions of documents and thousands of transactions daily. Ongoing efforts include onboarding new clients and migrating additional data from legacy systems.
The flexibility offered by their AWS solution permits scalability in line with increasing customer needs and paves the way for further enhancements, enabling rapid deployment of new features and global expansion. Collaboration with AWS has been beneficial, with regular consultations leading to best practice guidance and cost-saving recommendations. Challenges were promptly addressed, including minor modifications on the AWS side to enhance their services.
This transition has led to substantial savings on licensing, storage, and server infrastructure, with runtime service costs expected to decrease by over 95% due to the shift to serverless Lambda functions. Additionally, hosting APIs on AWS has improved performance and simplified new client onboarding, a process that can be streamlined further to facilitate client engagement.
For those returning to work after a break, this blog post on working moms transitioning back to work offers insightful tips. Additionally, if you’re looking for authoritative guidance on mandates regarding vaccines for city workers, consider visiting SHRM. For those interested in the hiring process at Amazon, Glassdoor provides an excellent resource.
Leave a Reply