arrow_back Back to Projects
Serverless RASTER AWS

RastLess

A serverless architecture for efficient visualization and analysis of large-scale spatio-temporal raster data in cloud environments.

calendar_today December 2022
code Status: In production

lightbulb Introduction

RastLess is a cloud computing innovation I developed at EOMAP for processing and visualizing large volumes of Earth observation raster data. The project addresses a fundamental challenge in the growing Earth observation market: how to efficiently process massive amounts of satellite imagery without the prohibitive costs of traditional server infrastructure.

In December 2022, RastLess was recognized as a double winner at the Space Awards in Rome during the Copernicus Masters celebration. The innovation won both the "Regional Prize Bavaria" by AZO and the "Frontier Tech Challenge" by the European Space Agency ESA, validating its potential to transform how Earth observation data is processed in the cloud.

crisis_alert The Challenge

For the growing Earth observation market, efficient processing of large raster data volumes in various resolutions is essential. As satellite imagery becomes more accessible and datasets grow larger, the traditional approach of running servers continuously becomes increasingly expensive and inefficient.

As I explained during the competition: "Continuously running servers are one of the most expensive cloud workloads, and low performance in accessing multiple data sources a real pain for users." Organizations working with Earth observation data face a difficult choice—either maintain costly infrastructure that runs 24/7, or accept slow performance when processing large datasets.

The challenge was clear: find a way to process hundreds or thousands of satellite images quickly and efficiently, while dramatically reducing infrastructure costs. This meant rethinking the entire approach to how raster data is stored, accessed, and processed in the cloud.

engineering The Innovation

In contrast to existing Data Cube applications, RastLess is a serverless solution for visualizing and analyzing distributed raster data. Instead of maintaining servers that run continuously, the system only uses computing resources when they're actually. This fundamentally changes the cost structure from fixed to variable, directly tied to actual usage.

Key Capabilities

Massive parallel processing

RastLess can process hundreds or thousands of satellite images simultaneously within seconds, enabling rapid analysis of large Earth observation datasets.

Flexible data handling

The system handles all sorts of projections and resolutions, working seamlessly with distributed data sources without requiring extensive preprocessing.

Cost efficiency

By eliminating always-on servers, RastLess dramatically cuts infrastructure costs while maintaining high performance for data access and analysis.

Improved user experience

Fast access to multiple data sources solves the performance pain points that have traditionally plagued Earth observation data platforms.

The development of RastLess was supported by Professor Adam Roe of CODE University Berlin and EOMAP's software team. Together, we focused on solving the real pain points of analyzing raster geo-data through a completely reimagined approach to cloud computing architecture.

trending_up Impact and Results

Seconds
Process hundreds or thousands of images
Flexible
All projections and resolutions supported
Reduced
Infrastructure costs dramatically lowered

RastLess demonstrates that serverless architectures can effectively address the unique challenges of Earth observation data processing. By eliminating the need for continuously running servers, the system provides a new approach to managing the growing volumes of satellite imagery while keeping operational costs under control.

The innovation has particular significance for the Copernicus program and the broader Earth observation community. As datasets continue to grow and more organizations seek to leverage satellite imagery, solutions like RastLess offer a path forward that balances performance, flexibility, and cost efficiency.

The recognition from ESA and AZO at the Copernicus Masters validates the potential of this approach to contribute to the future of Earth observation data infrastructure. The system continues to be developed and refined at EOMAP, serving real-world applications and informing best practices for serverless geospatial data processing.

emoji_events Copernicus Masters 2022

On December 1st, 2022, at the Space Awards in Rome, RastLess was recognized as a double winner during the Copernicus Masters celebration. The innovation received two prestigious awards, highlighting its potential to transform Earth observation data processing in cloud environments.

Regional Prize Bavaria

Awarded by AZO

Recognition for outstanding innovation in the Bavarian space and Earth observation sector, acknowledging RastLess's contribution to advancing cloud computing technologies for geospatial applications.

Frontier Tech Challenge

Awarded by the European Space Agency (ESA)

ESA's recognition of groundbreaking technological approaches in Earth observation, highlighting the innovation's potential to improve Copernicus cloud productivity and reduce infrastructure costs.

"EOMAP demonstrated a unique proposition to introduce new ways of EO data processing using serverless architectures. This can result in cutting infrastructure costs and improving Copernicus cloud productivity."

— Anna Burzykowska, Copernicus Innovation Officer, European Space Agency

play_circle Cloud Native Geospatial 2022

I presented RastLess at the Cloud Native Geospatial event in 2022, introducing the concept and demonstrating how serverless architectures can be applied to Earth observation data processing challenges.

groups Project Background

RastLess was developed during my time as a software engineer at EOMAP, a company specializing in Earth observation and water quality monitoring. The project emerged from real-world needs in processing large-scale satellite imagery for environmental monitoring applications.

Development Team

  • Marcel Siegmann — Lead developer and innovator, software engineer at EOMAP
  • Professor Adam Roe — Academic advisor from CODE University Berlin
  • EOMAP Software Team — Supporting development and integration

The collaboration between industry expertise at EOMAP and academic guidance from CODE University Berlin created an environment where innovative approaches to longstanding infrastructure challenges could be explored and validated in production settings.

Questions or Collaboration

For questions about this work or to discuss similar challenges in geospatial data processing, feel free to reach out.