Hexagon pilots AWS as the first step to creating Leica HxMap cloud-based software-as-a-service offering
In the geospatial profession, it’s easy to think our jobs revolve solely around imaging cameras, LiDAR sensors, and aircrafts. While these shiny pieces of hardware can be impressive and undeniably play the lead role in collecting the data that is the lifeblood of our industry, we often forget about the equally critical back-end workflows that turn this raw data into valuable information for our companies and our clients.
The truth is we are also in the big data business. Geospatial relies on high-performance computing (HPC) and big data processing as any other data-centric industry in the world today. The back-end networks, data storage devices, and compute nodes that perform the processing are every bit as vital to our project success as sensors and aircraft.
At Hexagon, we pay equal attention to both the front and back ends of geospatial projects. This was a major driver in the creation of the Leica HxMap multi-sensor software platform, that streamlines the processing workflow for all Leica Geosystems airborne sensors. But in the spirit of continuous improvements, we have piloted a cloud processing and storage project with AWS (Amazon Web Services). We expect this project to be the first step in offering HxMap in a cloud-based software-as-a-service configuration.
Leica HxMap On Demand
From a single user interface, HxMap’s suite of production tools processes raw datasets from quality control through to radiometrically and geometrically corrected 2D image and 3D point cloud-based products. Outputs are ready for stand-alone use, or further processing by third-party GIS applications and streaming platforms, or within HxDR, Hexagon's cloud-based storage platform for reality capture and geospatial data.
Customers benefit from HxMap multi-sensor workflow to process imaging and LiDAR data captured with any Leica Geosystems airborne sensor. Even if they own multiple sensors, they only require a single software, which saves costs and enhances efficiency. Personnel need only learn one simple user interface, and don’t waste time and avoid error as they transfer data from one application to another.
Additional benefits of HxMap include:
•Increased flexibility and scalability, from laptop to large HPC cluster• Focus on production, efficiency, and massive data throughput
• Merging 2D and 3D production generation workflows
• Open interface to other workflows using scripting and SDKs, allowing for easy adaptability to customer specific needs
Although Hexagon provides guidance on setting up an on-premises HPC infrastructure with specific customised recommendations, building an on-premises processing environment presents several challenges. The first, of course, is cost. Purchasing HPC equipment is a significant capital expenditure, and the investment doesn’t stop with the hardware. Configuring and continuously optimising the network, storage and nodes is a constant balancing act that necessitates the hiring of experienced IT staff.
The next issue is scalability — a seemingly never-ending challenge. At a minimum, both image and LiDAR data sets keep getting larger, which means the HPC environment that was perfectly balanced for optimal performance last flying season will likely need to be upgraded with new storage or nodes this year.
But an even bigger problem is that nearly every large client we work with has experienced the need for short-term expanded capacity to scale up a big or unexpected project. This is especially true in disaster mapping situations, where fast turnaround requires larger computer, storage, and network capabilities. Investing in additional infrastructure for a one-off project isn’t cost effective.
These two challenges — cost and scalability — are the main drivers for Hexagon’s pilot with AWS for cloud deployment of HxMap, with plans to create a software as a service offering in the near future. For many customers, the ability to pay for processing capabilities by the hour or by data set size will eliminate the need for investment in HPC hardware of their own. And the beauty of cloud processing is that users can easily scale their processing workflows up or down as needed, paying only for what they use at any given time.
Why AWS?
The success of running HxMap on AWS is not in question. We already know from our own experiences in the HxGN Content Program, and those of our customers are aware that the processing software works in several commercial cloud services. The goal of the pilot is to standardise the HxMap workflow in AWS so that it performs essentially the same for all users with predictable results at well-defined price points.
We chose AWS for the hosting of HxMap for three reasons:
1. Its lower price compared to competitors — AWS provides an excellent variety of services and tools at affordable costs. It provides an excellent variety of services and tools at lower pricing, which in turn, will make our software-as-a-service offerings more attractive as well.2. It is relatively easy to automate Windows clusters — Although HxMap runs on both Windows and Linux operating systems, Windows is the superior option for our geospatial customers because many downstream GIS, mapping and information extraction applications are Windows based. This means that most geospatial applications — including home-built ones — that customers use now in their post-HxMap downstream workflows can still run in the AWS cloud. In other words, complex workflows that our customers have built with HxMap and relied upon for years can be ported seamlessly to the AWS cloud in their existing configurations.
3. Many of our customers store and host final value-added products in AWS. Processing our data on the same cloud platform eliminates the need for data transfer costs from one cloud provider or even one region to another.
AWS Pilot Results
The early phase of the HxMap-AWS pilot using HxGN Content Program data has been completed successfully. Through this project, Hexagon is set to enhance the scalability, efficiency, and flexibility of the HxMap software, thereby contributing to streamline data processing for the Leica Geosystems’ airborne sensors.
