FarEye Saves $1 Million Annually with Compute Savings Plan, Amazon EC2 Spot Instances, and AWS Graviton Processors

With over 150 corporate customers in 30 countries and 5 global offices, FarEye is a last-mile technology partner for companies obsessed with delivery experience for their customers. Its vision is to ensure every delivery in the world reaches its destination punctually, efficiently, and sustainably. FarEye’s AI-powered platform transforms deliveries into a competitive advantage by empowering carriers and shippers to simplify complex last-mile logistics through orchestration, real-time visibility, and personalized customer experiences. The platform helps businesses improve consumer loyalty and satisfaction, reduce costs, and improve operational efficiency.

Ecommerce has changed the world's shopping habits, with consumers increasingly buying online for home delivery. This trend has led to significant investment in platforms to manage the order-to-door delivery process. However, with last-mile logistics consuming more than half of total costs, businesses use FarEye’s software-as-a-service (SaaS) web platform to optimize operations at this final stage. Gaurav Srivastava, cofounder and chief product and technology officer at FarEye, says, “Demand for our platform is growing by over 40 percent year-over-year as organizations try to make last-mile delivery work despite their tight margins.”

To maintain its edge in a competitive marketplace, FarEye focuses on product development. As the business expanded, its SaaS solution was seeing higher seasonal peaks and more extreme daily fluctuations in traffic. Comments Srivastava, “The challenge is balancing platform enhancement with cost, so our software protects customers’ profitability.” To ensure the infrastructure behind its platform could continue scaling, FarEye worked with Amazon Web Services (AWS)—which has supported the FarEye platform since day one. Srivastava explains, “AWS has consistently worked with us on platform optimization to reduce costs, improve performance, and better serve our clients.”

AWS proposed running FarEye’s Amazon Elastic Compute Cloud (Amazon EC2) workloads on Amazon EC2 Spot Instances, which provide steep discounts compared to on-demand prices for unused compute capacity on AWS. “The AWS team saw that we were consuming significant computing resources on integrations with carriers and suggested Amazon EC2 Spot Instances for greater value. Spot instances could tolerate interruptions or be scheduled across different timeframes, perfectly aligning with the demands of these integrations,” Srivastava comments.

FarEye, which is also an AWS Partner, adopted best practices for using Amazon EC2 Spot Instances to leverage lower costs, enhance application resilience, and minimize changes to its applications. The key changes the company made included giving each carrier integration its own virtual dockerized environment on Amazon EC2 Spot Instances. “We wanted to ensure possible disruptions were contained and didn’t impact multiple customers,” recalls Srivastava.

Following continued discussions with AWS, FarEye further optimized its platform’s infrastructure by moving the instances onto AWS Graviton Processors. The processors support a broad range of cloud workloads at a lower cost than x86-based Amazon EC2 instances. Within 12 months, FarEye migrated most of its On-Demand Instances to AWS Graviton Processors. For its Amazon EC2 Spot Instances, the company plans to diversify usage across multiple architectures, including x86. Srivastava confirms, “We have over 500 Amazon EC2 Spot and On-Demand Instances supporting workloads across our platform, with the majority utilizing AWS Graviton Processors.”

Combining the new instances and processors helped FarEye reduce its AWS compute costs by 65 percent. Says Srivastava, “Because final-mile delivery operates on such small margins, any savings we make helps us provide services to customers at a more competitive price, which is great for them and for our business growth.”

By moving to Amazon EC2 Spot Instances and AWS Graviton Processors, FarEye is saving $1 million yearly on compute expenses and can scale its platform for expansion and service development without worrying about the impact on expenditure. Baldeep Singh Kwatra, senior engineering manager at FarEye, says, "We achieved a seamless and error-free transition that reduced our AWS costs by 30 percent, improving performance by 10 percent and optimizing resource allocation for further investment in innovation."

Besides using the savings to give customers a compelling price point, FarEye is reinvesting in its technology. AI is rapidly enhancing last-mile delivery services, improving route allocation, and boosting capacity planning. FarEye has already added these types of AI capabilities to its platform and plans to deliver more. “Even though we've been recognized as a leader in AI-driven last-mile delivery management, we need to keep innovating,” says Srivastava. “And AWS is accelerating our adoption of generative AI to achieve this.”

The product development team uses Amazon Bedrock to experiment with multiple large language models for AI use cases. These include dynamic customer dashboards that provide real-time data visualizations of delivery operations and automated resolution of customer support tickets. Comments Srivastava, “With Amazon Bedrock, we’re seeking out practical generative AI–based solutions for a platform that can deliver new levels of value to our customers.”

For the next step in its optimization journey with AWS, FarEye is migrating its analytics data from a third-party platform to Amazon Redshift. The company also plans to use Karpenter to enhance node startup time and improve application availability. “We’re expanding our use of AWS services to ensure our IT aligns with our business strategy of providing greater customer value,” Srivastava concludes.