Feature Cache: High Performance Feature Cache with NoSQL for ArcGIS Server

—Alaor Bianco, Robson Junior, and Thales Oliveira (Imagem)

Performance and scalability demands have been increasing in the GIS world, and to address this, Esri has made great advances on image-based caches. That approach has some pitfalls, like lack of symbology customization and attributes retrieval on client side; even though ArcGIS Server dynamic services provide these functionalities, they are commonly an architecture bottleneck. Amid this challenging scenario, we have developed a solution that contains fast feature retrieval from a two-level cache:

• L1: InMemory Persistable (using RTree and QuadTree Algorithms)
• L2: MongoDB, a NoSQL database (with built-in spatial search and scaling out capabilities)

Feature Cache Service is a WCF REST service that returns features in GeoJSON format. The demonstration will show the service consumption by a GeoJSON layer on ArcGIS API for Silverlight.