Together with Softlayer Snowciety gave a presentation at GOTO Amsterdam 2013 about building custom maps using OpenStreetMap and SRTM data, POstgis/PostgreSQL as a datsbase, Mapnik as a renderer, Tilestance and Apache as the http servers and Leaflet as the javascript client.

6. Building maps for
apps in the cloud

8. Share your skiing with the world.
Join the Snowciety

9. Ski tracker

10. Friend finder

11. Share your skiing

12. Google Maps alternative
because they charge heavy users nowadays

17. Freedom
create radical designs

20. It is all Open Source
even the data from NASA

21. data database renderer http server

22. client

23. OpenStreetMap
wikipedia for maps

24. SRTM
Shuttle Radar Topography Mission

25. Lots of data
our cluster contains roughly 3,5 terabyte of unrendered data

26. 350 GB of OSM
PostgreSQL with PostGIS plugin

27. 3 TB of GEOTIFF

29. Mapnik
renderer

31. §

44. Journey to get here

47. Journey to get here

48. §

49. png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
png
256x256
png png png png png png png png png png

50. Not only server-side
also client-side library to interact with map

51. An Open-Source JavaScript Library
for Mobile-Friendly Interactive Maps

52. Apache httpd
No, not really the chopper

53. Tile Stache
Translating Apache requests into Mapnik render commands

54. Mapnik
the renderer from before

55. Caching
Memcache, Disk, S3

56. Caching
Memcache, Disk, S3

58. Importing data
takes a loooooooooooooooooooooooooooooooooooong time

59. 25GB
gzipped xml file

60. and you know how well plain
text can be gzipped
file is roughly 250GB unzipped

61. 3 weeks and counting :-|
M1 Large Instance
7.5 GiB of memory, 4 EC2 Compute Units

62. High-Memory Quadruple Extra
Large Instance
68.4 GB of memory, 26 EC2 Compute Units, 24 EBS blocks RAID 1
$1.640 per Hour = $ 1180 / month
29 hours!

63. Dedicated hardware
68 GB of memory, 8 Cores, 1 SATA disk
$ 700 / month
10 hours

64. Macbook Pro
16 GB of Memory, SSD

65. 8.5 hours
it is all about lots of memory and lots and lots of IO speed

66. and that was just OSM!

67. SRTM
Shuttle Radar Topography Mission

69. about a minute
per 1” x 1” file

70. from -180,0 to 180,90
28.800 files

71. 20 days later
we were done importing

73. So now you know
how it is made

74. Future?
More sources, better maps, and..

75. Vector maps!

81. How do I get started?
Mapbox.com
download TileMill

82. Building your on OSM server (incl all build commands)
http://weait.com/content/build-your-own-openstreetmap-server-lucid
Working with terrain data (hillshading, slopeshading, color-relief)
http://www.mapbox.com/tilemill/docs/guides/terrain-data/
for when viewing this presentation on Slideshare
Tutorials

83. THANK YOU
@timanrebel

84. Rate me!
GOTO Guide App
!