Case study: A full-fledged cutting-edge FOSS4G map production system

Video thumbnail (Frame 0) Video thumbnail (Frame 1380) Video thumbnail (Frame 1956) Video thumbnail (Frame 5896) Video thumbnail (Frame 11049) Video thumbnail (Frame 19265) Video thumbnail (Frame 20093) Video thumbnail (Frame 21748) Video thumbnail (Frame 22406) Video thumbnail (Frame 23271) Video thumbnail (Frame 28548) Video thumbnail (Frame 30768) Video thumbnail (Frame 36410)
Video in TIB AV-Portal: Case study: A full-fledged cutting-edge FOSS4G map production system

Formal Metadata

Case study: A full-fledged cutting-edge FOSS4G map production system
Title of Series
CC Attribution - NonCommercial - ShareAlike 3.0 Germany:
You are free to use, adapt and copy, distribute and transmit the work or content in adapted or unchanged form for any legal and non-commercial purpose as long as the work is attributed to the author in the manner specified by the author or licensor and the work or content is shared also in adapted form only under the conditions of this license.
Release Date
Production Year
Production Place
Seoul, South Korea

Content Metadata

Subject Area
The development and the usage of National Land Survey of Finland's dynamic and high performance map production system is described in this presentation. The system is currently in use and serves map images both to customers and to NLSFI production systems. The data in the map production system are open data and being updated on a weekly basis. When the data get updated, a RSS-feed is generated. Based on the feed, the map products are updated. Data is stored, updated and replicated in PostGIS. Map pictures are rendered in GeoServer. The visualization of the maps is based on SLD-stylesheets. SLD-stylesheets enable the same data to be visualized in several different ways. GeoServer in conjunction with SLD-stylesheets offers a Web Map Service (WMS). Map images are delivered via a high performance MapCache Web Map Tile Service (WMTS) and as image files via NLSFI download service. The system is designed to be expandable and is currently being further developed to enable the pro-duction of on-demand printed maps.
Presentation of a group Computer animation Strategy game Mapping Open source Decision theory Self-organization System programming Open set Product (business)
Presentation of a group Computer animation Mapping System programming Product (business)
Server (computing) Texture mapping Mapping Multiplication sign Projective plane Bit Database Open set Mereology Product (business) Cache (computing) Web service Goodness of fit Computer animation Software Videoconferencing MiniDisc System programming Self-organization Thermal conductivity Computer architecture
Slide rule Server (computing) Texture mapping Mapping Computer file Real number Structural load Bit Database Insertion loss Open set Mereology Product (business) 4 (number) Medical imaging Web service Mathematics Process (computing) Computer animation Internetworking Order (biology) Lastteilung Cycle (graph theory) Computer architecture
Point (geometry) Metre Slide rule Server (computing) Pixel Greatest element Computer file Decision theory Multiplication sign Workstation <Musikinstrument> 1 (number) Set (mathematics) Solid geometry Open set Product (business) Number Medical imaging Web service Different (Kate Ryan album) Square number Audiovisualisierung Energy level Associative property Area Texture mapping Mapping Cellular automaton Moment (mathematics) Electronic mailing list Bit Process (computing) Computer animation Personal digital assistant Self-organization Musical ensemble Reading (process) Resultant
Medical imaging Word Computer animation Computer file Personal digital assistant Multiplication sign Single-precision floating-point format Order (biology) Representation (politics) Electronic visual display Database Insertion loss
Computer animation Personal digital assistant
Goodness of fit Computer animation Visualization (computer graphics) State of matter Database Position operator
Point (geometry) Functional (mathematics) Building Dynamical system Computer file State of matter Real number Multiplication sign View (database) System administrator Workstation <Musikinstrument> 1 (number) Mereology Field (computer science) Web service Mathematics Different (Kate Ryan album) Extension (kinesiology) Posterior probability Vulnerability (computing) Area Distribution (mathematics) Texture mapping Mapping Interface (computing) Projective plane Moment (mathematics) Planning Bit Line (geometry) Computer animation Order (biology) Right angle
Area Server (computing) Functional (mathematics) Distribution (mathematics) Computer file Mapping Interface (computing) Workstation <Musikinstrument> Field (computer science) Product (business) Human migration Web service Mathematics Process (computing) Vector space Lecture/Conference Tower Personal digital assistant Different (Kate Ryan album) Table (information) Position operator Probability density function Form (programming)
but I already welcome all this session my
name's circumvent in from I'm from the National Land Survey of Finland and never a GIS specialist during a specialist there I think it ight interviews myself now because I most of the 1st presenter the uh I am also the chairman for decision so thank you to the organizers organizers for this honour pity the topic of this session is working efficiency open source GIS and we will have 4 presentations the session data the resentations decorum map production decision-making systems and open strategies so the I was not government that my own presentation it is the topic
of my present of my presentation is they cutting edge for gene map production system and the topics I cholera in
this entation is to have them to a bit about the background why we started to produce or develop a new system and will look at the time on the architecture of the production system we will go through in briefly used technologies the city what we have accomplished and look at some of the challenges we have encountered sold during the way then we'll see how we further are going to develop this this production is map production systems so about the background the 1st reason why we started to think about to develop a new map production system what the the had the VMS services what a few years ago 4 5 years ago the increase usage started to increase a lot so they can take Lord more Soviet needed quickly to make that w MDS Cache servers and and the reason was that the decided to use open software in or organization and existing software that we used to produce maps the lifespan of doing that systematic complement and that indeed the server up-to-date maps and also take part in international projects and that this needs would not be fulfilled easily with the old system so a couple of good reasons why we should start to develop a new a new system just said the 1st thing to be done was a WTA style caching without interrupting and they start to 1st 1st based for the MAP conduction system in 2012 the so what the architecture and this time finally managed to make a simple architecture picture so it's basically we have PostGIS database the video so that mapping service and we use map catch as that we in theUS potash and then we have uh or assess feeds uh all reach the about and the database discusses which with each other will come back more today's and
then subscripting to keep it all together the so if you are a bit more in detail look at the picture it's basically the same as the
previous slide but a bit more in detail production database where uh real estate engineers and cartographers make changes if is a new building the Torah Old made some there somebody digitizes it and it in the production database and the changes will each day be transferred to the father load service and then we have loading server on which forced yes France it discusses with this file download service for Atom feeds and downloads changed beta sheets the loading server lost in the post js and at each week and on Sunday we database become from the loading server and transfer do over map servers and on this map servers uh we have just over running the VMS and it an exact copy of post GIS database that is on the loading loading server so the the copies sex act exactly the same act Sunday morning but the Monday of course the production starts again and then the make changes and models in the 4th year so it's a week weekly cycle this in front of the map service we have a load balancer and behind that you have maps over WTA sketches and then for each week we know what has been changed in the database and as I said we use use RSS feeds so we read the change various and populate those once again the WMT a sketch and it the same with a map image files we have we also sort the data as map image files have a look at those what they are in the next slide on all the following and those loss of interest with the lower the Open Data fondled service and finally the arrows from top describes internet for users that then again the use of the MAXserver wired the same load balancer so this is basically a simplified picture of our architecture so let's look a bit more on the different different parts of the of the service the 1st thing we are going to look at this daughter then we have an in-house middle of yellow order that takes care of the extracted from a long process and the what you say about voltages yeah not think on 3rd
any problems maybe not in a promise and at all but not any real problems verbs very well for our needs and then
about the they map server running on just over the 2 . 6 . 2 at the moment it has all the features we need but of course they're going to updated at some time the use of a processing services for some kind of day that generalization of some skates and and they're they're just some numbers to give you an some kind of picture what kind of data the have working with you have to find a distorted the generalized topographic data from my roughly 1 to 10 thousand up to 1 2 4 . 5 million so we have the different dot sets in the National Land Survey of Finland all already generalize that you can use for my production almost 400 as a list of this of these buys the unreported 2 layers 52 labeled so it's a quite a few layers and and as of this stars the the to publish and about the performances we run these wasn't too physical servers at the moment and then a 4 CPU and 256 be about will save further on in the temptation is c uh what that means or what kind what we what we can achieve with those ones and about the the and service so that I catch you as maybe somebody is already wondering why under the map Cashman use just over on the 1st slide and I said that the 1st decision to be made was to establish podcasts and at that point he did the decision did the decided to use MAP guassian uh there was some reasons that they are at speed they should be quite agree like quite similar or no big differences so we just found out that my precious of B 2 to better suited for our needs because In built you delete the preceding it's easy to configure and alternate and for our needs the general features of a little bit what did what can what could be said about that is that the also these to work very well together so that's that's something to keep in mind too and about their performance seemed area Finland which is about 300 square me this with more takes roughly 20 always into 2 visualizations and that the consumer levels on which the most accurate is 1 meter repeats and this results in about 10 million pounds of association and 60 gigabytes of 25 gigabytes the depending on the mission station will look at the 2 2 different kind of visualizations in the next slide and read updates takes about 1 hour the this micellization Toronto but solid achievements what we have accomplished each week the changed areas are updated of India's by cash and we have a bald and and in the previous slide I was talking about 2 additional stations but that was just to give you some numbers we have about 10 products that are continuously updated to of this service and they are updated up to 1 meter per pixel solution and a theUS cash has about 2 6 2 million hits the each mom that is no accurate number 1 to give it up by the what kind of usage you have and each week also to cross the fossils changed areas are updated and transferred to Open Data knowledge service so at the bottom you can see a few different kind of visualizations of the data you see that 3 once on the left from you are quite similar and then the back room up so these are different versions of for topographic map both for use all outside and inside of organization and the but then you have a background map with withdrawal next the we looked at a bit of what what the image files or in the in the case of the topographic maps they are image fast the thought 12 kilometres bias 12 kilometres and the solution is 1 meter the big cell had is 1 this risation that look like this which is freely available from all open file download service and and then
we have another 1 that looks like this as you can see this is a fully uh non and the alized map that music has no and and analyzed cells it just has the little ones which is the heart of this specified there was a need to have a map like this and that reflect these 2 so so they're supporting that also but that
representation but a few words about the challenges we have encountered the biggest loss has been induced over and this is the order styling of data 1st of all it's so hard to render all labels and probe names and it's harder and they're all labels but many also the end the old names it's it's even trickier and it this of course of the fact that when you request and Borland books on containing the image if affair the label uh reaches the file with the time stops we don't know if it should be the end of the and then this is a pretty advanced topic I cannot explain it altogether but we ended up with a solution where each label it is is loaded into the database as single characters so at these pictures here is there's no labels but for example
displays name lock your this is 1 feature is 1 feature case 1 feature so when we now come to the border like
this you all in normal cases would not be sure but in just a it's a
the days of an eruption Potchefstroom rituals laughter draw de label on the border and again do now when no later
comes from the left side yeah like it's the position of fail is down here in the left corner if ever to be here you have no idea if it should be drawn on not but then again today the BMS repressed you can add a buffer so that the data being granted from the database is slightly bigger than the picture of the state so using those 2 the tricks we could manage to get all the the place names and also get a good visualization of role names
and there's something it on our maps that the could call so all area fields that is what you can see on the picture below of the green 1 in fact the this year because mouse there germ metaphor this line is just green line but then we needed to have In the area of Finland inside of it so what we need to do what if you have a materialized view of used in this in posteriors very 1st offset the line and then buffering and the buffered as being created is then ned would something in in a normal way and quite the same thing you want to for some roads we want to have the label on top of the order of the the road but not not to be on the road so that also is naturalized to you in in both GIS which is made up of a function and the and the 1 challenge that is still there is that you have a lot of facility files and then this from unknown administrators point of view a bit of a challenge to keep them up to date and to applies some change that should take they did it should be so much but not in all of them so as the last slide and tell you something about the work in this the European right now and on this service and what the other thing to do in the future right now we are at building on extension or interface to this service from which he can request different kind of offer printed maps or maps that are suitable for printing that means that the optimal quite big 0 uh the graphic maps but also a different kind of fat maps that are used in the house for real estate uh purposes for example it is there are also taking part in international projects in Europe that means if you left project them into by project as you can see it also plans for the future of the left and inspire projects continuing so those ones the are and you have taken Boston as a planning to To make a dynamic WMS service which means it is to be through dynamic in that way that the user can all states of the files affect the distribution station and also the data be shown the map picture he gets and 1 demand at the heaviest at the want to at least the real estate data to be updated everything in 4 hours so or even less than 24 hours but during the nite time so much morning when business always starts they would have well yesterday's situation online in also in a lot of the NPs service so thank
you this was the end of my expectation of foreign more information please feel to contact me and then is the how any Christians this the take microphone yeah if the like just the it is Paikea to upgrade from 2 . 6 2 because there's no security vulnerability that's will publish because we want to watch for it if you need to go at least 6 full her the yeah body it there since it's a running we are not exporting it there so it's not it's through the EU XML requests so you could pay half 0 yeah yeah I'm I'm I'm out there that at the moment of the book that preceded it then they will update for other reasons to 1 now 1 comment and 1 question the comment is uh you this said had difficulties with labels which where on the heavier the heavier tried obtaining in Jerusalem yeah and then work for you that's good question i and
it comes back to do today thing it's hotter and all labels and role names could you do matter filing the old names will be at the borders sooner or later and if they are not that we did not do not have a fixed position in our case get can decide the positions of 2 interrelated goes along the because and our migration is that's you also said the euro doing a partial updates although all the generated files is the the update when you're over the missing have changes in Europe they they generated towers you only have they'd so it's the partially not completely yeah OK how do you how do you identify how the year of those this process of process workers this those standard functionality you did some appropriated solution to that at what to do is that the in posting is we have a table for each each for each table we have a field that is don't change date the date when it has been updated the last but then Avagio or assisted services through which we can ask for a product for example for this map productive there's topographic map of mortars I'm showing what form for each map you can ask which areas have been changed since last week and into each de de table field and change the date and returned to those data sheets to as an XML files to map it to to the server to map this is running and and we know what all the data what as commented that he could have distribution station is not hard and you have a vector service you should consider not to not to update them by but to just delete them and wait for a user to request the area and then at the whole populated using them has warning when is yeah workflow for producing can printed maps so that that it's the same styling is for the web and plus all the tent like things around them is the I didn't tell everything about it but the virtual was basically that we are building an interface that is set separate from from this service but if 3 Christian maps while at the billionaires from just arrest PDF files in different kind of layers and in our job of let's just passed different kind of layers together and they get PDF map but if you are interested in it we can discuss further on it deems ninteen OK thank you for the what was changed thanks thank you and then as next I would like to welcome the me lamb and ball of itch the most similar I have another could not make it today unfortunately what meal will again at this up