What we do
Popular web mapping platforms, such as Google Maps, Bing Maps and MapQuest have sparked public interest in mapping and spatial visualization.
We believe these mapping platforms can do even more.
We are extending the reach of these platforms into areas of real-time visualization of large volumes of spatial data, business intelligence, and other domains that need high-performance mapping.
How do we do it?
Our pursuit relies on the careful combination of these technologies and services:
- Spatial Databox: A scalable, low-latency web service that delivers location-based content to map-based client applications.
- Google Maps API for Flash: It's like the usual Google Maps that we all know and love, but is built on a different, turbocharged platform.
- ActionScript 3: An advanced browser-side programming language that simplifies the development of correct, resilient applications.
- ActionScript Virtual Machine 2: Accelerates web application execution to deliver near native-level performance. (Also known as AVM2.)
- Amazon EC2: An economical way to serve large spatial data sets to a large number of client applications.
- Google App Engine / Java: The easiest and most economical way to serve smaller spatial data sets to a continually varying number of client applications. (Especially capable of handling unexpected surges in traffic.)
Users quickly abandon slow or erratic websites. So, as we apply these technologies, we obsessively pinpoint and eliminate overhead and inefficiencies, wherever they may occur:
- We deploy on the appropriate cloud infrastructure, to balance cost and mitigate demand surges. (For us, the equation is not Amazon or Google, but Amazon and Google.)
- We write web services in Java and steer clear of unnecessary synchronization, buffer copying and heap utilization.
- We minimize application latency by crafting tight, bespoke application frameworks that load quickly and render content progressively. (Progress indicators and hourglasses are a no-no!)
- We tie web service and application together, through proprietary binary protocols. Our binary protocols also reduces network bandwidth and server overhead (XML and JSON are used sparingly.)
The following sections touch on the need for high-performance mapping, and related events in industry.
