Impact of user concurrency in commonly used OGC map server implementations

29
INFOCOMP 2011 Impact of User Concurrency in Commonly Used OGC Map Server Implementations INFOCOMP 2011 October 2011 Used OGC Map Server Implementations Joan Masó, Paula Díaz, Xavier Pons, José L. Monteagudo-Pereira, Joan Serra-Sagristà, Francesc Aulí-Llinàs Center of Research in Ecology and Forestry Applications Universitat Autònoma de Barcelona

description

Masó, J., Díaz, P., Pons, X., Monteagudo, J.L., Serra, J., Aulí, F., (2011). Impact of user concurrency in commonly used OGC map server implementations, en: Proceedings of INFOCOMP. Barcelona, October 2011. ISBN: 978-1-61208-161-8.

Transcript of Impact of user concurrency in commonly used OGC map server implementations

Page 1: Impact of user concurrency in commonly used OGC map server implementations

INFOCOMP 2011

Impact of User Concurrency in Commonly

Used OGC Map Server Implementations

INFOCOMP 2011October 2011

Used OGC Map Server ImplementationsJoan Masó, Paula Díaz, Xavier Pons, José L. Monteagudo-Pereira,

Joan Serra-Sagristà, Francesc Aulí-Llinàs

Center of Research in Ecology and Forestry Applications

Universitat Autònoma de Barcelona

Page 2: Impact of user concurrency in commonly used OGC map server implementations

Introduction

INFOCOMP 2011October 2011

Introduction

Page 3: Impact of user concurrency in commonly used OGC map server implementations

Three Steps for Disaster Management

INFOCOMP 2011October 2011

GEO-PICTURES is an EC FP7 SPACE project with the aim of

integrating satellite imagery with in-situ sensors and geo-tagged images as a tool for decision making in emergency crisis

Page 4: Impact of user concurrency in commonly used OGC map server implementations

Rapid Mapping and Technology

INFOCOMP 2011October 2011

Page 5: Impact of user concurrency in commonly used OGC map server implementations

The images

� 22 satellite images of GeoEye-1 (Orthorectified GeoTIFF; provided by Google)

� (http://www.google.com/relief/haitiearthquake/geoeye.html)

� Covering Port-au-Prince

INFOCOMP 2011October 2011

� Covering Port-au-Prince and surroundings

� 16-01-2010, 3 days after the Earthquake

� Each image has 196 373 kb � 4.21 Gb

� 40 994x57 392 pixels

pdiaz4

Page 6: Impact of user concurrency in commonly used OGC map server implementations

Diapositiva 5

pdiaz4 Al Web de descàrrega posa:

By downloading these files, you agree to use the imagery solely for non-commercial use related to emergency relief, and to provide a proper and distinct photo credit to “GeoEye Satellite Image.”

Això significa que hem de posar el logo de GeoEye a la presentació?pdiaz; 13/10/2010

Page 7: Impact of user concurrency in commonly used OGC map server implementations

The service

� We are going to test implementations of two standards:

� Web Map Service (WMS) standard

INFOCOMP 2011October 2011

� Web Map Tile Service (WMTS) standard

� Both standards are Open Geospatial Consortium standards

� Assess performance

Page 8: Impact of user concurrency in commonly used OGC map server implementations

Web Map Service

INFOCOMP 2011October 2011

This is a map

Page 9: Impact of user concurrency in commonly used OGC map server implementations

Web Map Tile Service

INFOCOMP 2011October 2011

This is a tile

Page 10: Impact of user concurrency in commonly used OGC map server implementations

INFOCOMP 2011October 2011

Page 11: Impact of user concurrency in commonly used OGC map server implementations

Methodology

INFOCOMP 2011October 2011

Methodology

Page 12: Impact of user concurrency in commonly used OGC map server implementations

The analysisServers ClientsStandardsData

Web Map Service

(WMS)

Web Map Service

INFOCOMP 2011October 2011

Web Map Service

Cache (WMS-C)

Tile Map Service

(TMS)

Tile Map Tile Service (WMTS)

Page 13: Impact of user concurrency in commonly used OGC map server implementations

Traditional WMS server-client interaction

WMSServer

request GetMap

URL

INFOCOMP 2011October 2011

response

� All studied protocols request maps by creating an URL with specific syntax� http://www.ogc.uab.es/cgi-bin/SIGMA/MiraMon5_0.cgi?VERSION=1.1.0&

REQUEST=GetMap&SRS=EPSG:27573&BBOX=532776,22819,538776,26419&WIDTH=600&HEIGHT=360&LAYERS=mh-andorra&STYLES=&FORMAT=image/gif&TRANSPARENT=TRUE

� URL requests were randomly generated and sent from different clients

� The time response is stored in logs and latter analyzed

Page 14: Impact of user concurrency in commonly used OGC map server implementations

Results

INFOCOMP 2011October 2011

Results

Page 15: Impact of user concurrency in commonly used OGC map server implementations

Evaluation of WMS Concurrent Requests to a Single Server

WMSServer

reqGetMap

INFOCOMP 2011October 2011

WMS Server Server

res

Page 16: Impact of user concurrency in commonly used OGC map server implementations

Evaluation of WMS Concurrent Requests to a Single Server

�More than one hundred different requests were done (without optimizing speed configurations).

�The influence of the pixel size and the image size in

INFOCOMP 2011October 2011

�The influence of the pixel size and the image size in the time response were evaluated

�The requests were made from up to 5 concurrentclients.

�The time response for the requests are exposed in graphs.

Page 17: Impact of user concurrency in commonly used OGC map server implementations

Evaluation of WMS Concurrent Requests to a Single ServerResponse time of 5 different server vendors at different scales (pixel sizes) each one under

5 simultaneous requests

10.000

MapServer

GeoServerMiraMon Server

ArcGIS Server Express Server

INFOCOMP 2011October 2011

0.010

0.100

1.000

0.0001 0.0010 0.0100 0.1000 1.0000 10.0000 100.0000

Pixel Size (seconds of arc)

Tim

e (s

econ

ds)

Page 18: Impact of user concurrency in commonly used OGC map server implementations

Evaluation of a Cluster of Servers

�To overcome the performance degradation in

concurrent requests a possible solution is to set up a

cluster of servers

INFOCOMP 2011October 2011

cluster of servers

�The cluster of servers act as a virtual single server

� 6 computers are able to respond at same time to different

clients as if they were like a faster single server

Page 19: Impact of user concurrency in commonly used OGC map server implementations

Evaluation of a Cluster of ServersEvaluation of the response time for Pixel Size (Clients to MiraMon Single Server)

60.0

80.0

100.0

120.0

140.0

160.0

180.0

Tim

e (m

illis

econ

ds)

17 clients

14 Clients

11 Clients

8 Clients

4 Clients

1 Client

INFOCOMP 2011October 2011

Evaluation of the response time for Pixel Size (Clients to MiraMon Server Cluster)

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

160.0

180.0

0.0000 2.0000 4.0000 6.0000 8.0000 10.0000 12.0000 14.0000 16.0000

Pixel Size (seconds of arc)

Tim

e (m

illis

econ

ds)

17 clients

14 Clients

11 Clients

8 Clients

4 Clients

1 Client

0.0

20.0

40.0

0.0000 2.0000 4.0000 6.0000 8.0000 10.0000 12.0000 14.0000 16.0000

Pixel Size (seconds of arc)

Page 20: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response: Sequential

WMSServer

reqGetMap

reqGetMap

reqGetMap

reqGetMap

reqGetMap

INFOCOMP 2011October 2011

Server

Page 21: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response

� Some WMS clients are able to tile the space in a regular matrix of small pieces.

� They need several tiles to cover the whole viewport

� They can recycle some tiles when the user moves the view laterally

INFOCOMP 2011October 2011

They can recycle some tiles when the user moves the view laterally

� Also can take advantage of the cache mechanisms

� If the caching mechanism cannot help the response time can increase even if each tile is smaller that the whole view

� Tiled clients (tiles of 256x256 pixels) were simulated in three configurations.

Page 22: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response: Sequential�Results of the WMTS speed metrics

Time response for sequential 256x256 tiled requests on a pure WMS server

10MapServerGeoServerTilecache

INFOCOMP 2011October 2011

0.01

0.1

1

0.000

90.0

010

0.002

40.0

029

0.005

10.0

076

0.010

20.0

145

0.018

80.0

246

0.033

10.0

462

0.185

50.2

130

0.265

40.4

717

0.567

01.0

383

1.642

5

seconds of arc

Sec

on

ds

(tim

e)

TilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Sequential tiled WMS

Page 23: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response: Concurrent

WMSServer

reqGetMap 1reqGetMap 2reqGetMap 3reqGetMap 4reqGetMap 5

INFOCOMP 2011October 2011

Server

Page 24: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response:Concurrent�Results of the WMTS speed metrics

Time response for sequential 256x256 tiled requests on a pure WMS server

10MapServerGeoServerTilecache

Time response for unlimited concurrent 256x256 tiled requests on a pure WMS server

10MapServer

INFOCOMP 2011October 2011

0.01

0.1

1

0.000

90.0

010

0.002

40.0

029

0.005

10.0

076

0.010

20.0

145

0.018

80.0

246

0.033

10.0

462

0.185

50.2

130

0.265

40.4

717

0.567

01.0

383

1.642

5

seconds of arc

Sec

on

ds

(tim

e)

TilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Sequential tiled WMSConcurrent Tiled WMS

0.01

0.1

1

10

0.00

090.

0011

0.00

270.

0049

0.00

760.

0110

0.01

590.

0245

0.03

310.

0627

0.19

150.

2348

0.47

170.

5745

1.17

30

seconds of arc

Sec

on

ds

(tim

e)

GeoServerTilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Page 25: Impact of user concurrency in commonly used OGC map server implementations

Tiling the Request and the Response: Semi-concurrent

�Results of the WMTS speed metrics

Time response for sequential 256x256 tiled requests on a pure WMS server

10MapServerGeoServerTilecache

Time response for unlimited concurrent 256x256 tiled requests on a pure WMS server

10MapServer

Time response for up to 4 concurrent 256x256 tiled requests on a pure WMS server

INFOCOMP 2011October 2011

0.01

0.1

1

0.000

90.0

010

0.002

40.0

029

0.005

10.0

076

0.010

20.0

145

0.018

80.0

246

0.033

10.0

462

0.185

50.2

130

0.265

40.4

717

0.567

01.0

383

1.642

5

seconds of arc

Sec

on

ds

(tim

e)

TilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Sequential tiled WMSConcurrent Tiled WMS

0.01

0.1

1

10

0.00

090.

0011

0.00

270.

0049

0.00

760.

0110

0.01

590.

0245

0.03

310.

0627

0.19

150.

2348

0.47

170.

5745

1.17

30

seconds of arc

Sec

on

ds

(tim

e)

GeoServerTilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Semi-concurrent Tiled WMS

0.01

0.1

1

10

0.00

090.

0011

0.00

270.

0049

0.00

760.

0110

0.01

590.

0245

0.03

310.

0627

0.19

150.

2348

0.47

170.

5745

1.17

30

seconds of arc

Sec

on

ds

(tim

e)

MapServerGeoServerTilecacheMMServerArcGIS ServerExpress ServerGeoWebCache

Page 26: Impact of user concurrency in commonly used OGC map server implementations

Conclusions (1/2)

� The work presented covers:

� A metrics on WMS and WMTS services

� GeoServer, MapServer, MiraMon Map Server, ArcGIS Server, Express Server

� TileCache GeoWebCache

� A set of recommendations of Disaster Management

Easy to setup: MapServer

INFOCOMP 2011October 2011

� Easy to setup: MapServer

� Easiest configure and update: GeoServer

� Fastest: Express Server

� The speed tests described are a practical demonstration of the suitability of certain servers and service configurations in certain domains where reliability of services is imperative

� We have seen differences in performance of 2 order of magnitude.

� All the analyzed servers have slower performances when the number of simultaneous clients is increased

� A cluster of server dramatically improves performance

p2

Page 27: Impact of user concurrency in commonly used OGC map server implementations

Diapositiva 25

p2 HE FET DUES OPCIONS PER A LA DIAPO 1/2 DE CONCLUSIONS, UNA MÉS DENSA I L'ALTR AMOOLT MÉS LLEUGERA. TAMBÉ POTS COMBINAR LA LLEUGERA 1/2 AMB LA 2/2 QUE TROB QUE ÉS DENSA PERÒ QUE ÉS LA QUE VAREM FER A VENÈCIA. p.diaz; 25/10/2011

Page 28: Impact of user concurrency in commonly used OGC map server implementations

Conclusions (2/2)� In order to improve performance, some clients request tiles to servers that are not

prepared to serve them

� This results on no better performance in some servers

� Server optimization for tile requests is needed.

� Web clients auto-impose themselves a limit in the number of parallel request

INFOCOMP 2011October 2011

� Web clients auto-impose themselves a limit in the number of parallel request

� We saw that this more conservative strategy results on better performance

� MapServer and GeoServer with common open source services that do not require any data preparation process but

� their performance is worst than other services that require indexing methods like MiraMon Map Server

� MapServer (based on C++ code) performs better than GeoServer (based on Java code) under single client requests, but GeoServer is surprisingly faster under concurrent simultaneous requests.

Page 29: Impact of user concurrency in commonly used OGC map server implementations

Thank you!

INFOCOMP 2011October 2011

Thank you!

[email protected]