I have a fairly simple set of requirements, but I'm struggling to find anything that will handle them "out of the box" at scale. This could be my ignorance with GIS terms, or that there is nothing.

Essentially, we would like to load a system with thousands, or even millions of geographic "areas" (polygons / multi-polygons). A user will define an area/ line/point. We then need to see

a) what stored areas does the user geometry intersect with and

b) get the geometry defined by that intersection. The application is very much read heavy, with the stored "areas" changing infrequently and not in real time.

The SQL Server geographic data type seems to have this functionality (as does PostgreSQL) and would be a good solution for storing the data, but I worry about scale in terms of processing intersections with masses of data and many concurrent users (public facing website).

Cosmos and Mongo DBs seem to allow intersection detection, but not geometry generation.

I was expecting there would be some sort of a "server" (with many GPUs probably) that I could deploy across AWS and handle trillions of these sorts of operations per second, but so far my searches have come up empty. This is surprising because what I'm doing seems like a common use case.

  • You are on the right track with Postgres (Postgis). Postgres can be containerized. So it wouldn't be one instance handling all of the queries. If you are using AWS then it might be multiple AMIs but still these would scale up and down as demand grew and deflated. – enolan Aug 22 '18 at 15:31

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.