International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Volume XXXIX-B4, 2012 
XXII ISPRS Congress, 25 August - 01 September 2012, Melbourne, Australia 
419 
4.2.2 Case Study 2: Case study 2 tries to compute the length 
of railways in each of the administrative districts they pass 
through. Similar to case study 1, we use GML to represent the 
needed spatial data (railway layer) on the server side. We also 
represent the district boundary layer in SVG and deliver it to the 
browser side as the initial UI. The task is carried out by the 
following steps: 1) compute the railway segments in each 
district (using Intersection operator); 2) calculate the length of 
each railway segment (using Length operator); 3) sum up all the 
railway segments for each district; 4) filter out districts which 
have no railways. The SESQL sentences are listed in the 
Appendix. Figure 5 depicts the results. It lists the names of all 
relevant districts, and their lengths of railways in the listbox 
shown at the right-bottom comer. These districts are also 
highlighted in the map view. 
Figure 5. Listing the length of railways in each of the 
administrative districts they pass through 
We also compare the amount of transmitted data when using 
different solutions for accomplishing this task. Table 2 depicts 
the results, and shows that our proposed solution has a smaller 
network transmission load between server and browser sides. 
Server- 
side 
solution 
Client- 
side 
solution 
Layer- 
based 
solution 
The proposed 
solution 
Step 1 : 
Buffer 
28,493 
7,211 
28,493 (on 
server) 
7,211 (on 
browser) 
Step2: 
Length 
9,925 
0 
0 (on 
browser) 
0 (on 
browser) 
Step3: 
SUM 
648 
0 
0 (on 
browser) 
0 (on 
browser) 
Step4: 
Filter 
582 
0 
0 (on 
browser) 
0 (on 
browser) 
Total 
39,648 
7,211 
28,493 
7,211 
Table 2. Comparisons of the second case study (data amount is 
measured by Byte) 
4.2.3 Case Study 3: Case study 3 focuses on how land uses 
of each administrative district along railway “Guang-Mei-Shan” 
change between 1987 and 1996. Similarly, we use GML to 
represent the needed spatial data (railway and district-centre 
layers) on the server side. We also represent the district 
boundary layer in SVG and deliver it to the browser side as the 
initial UI. 
The task is carried out by the following steps: 1) calculate a 20 
km buffer of railway “Guang-Mei-Shan” (using Buffer 
operator); 2) identify the districts whose centres are located in 
this buffer (using Within operator); 3) Use the statistics function 
to generate the bar graphs of changes of land uses for every 
identified district. The SESQL sentences are listed in the 
Appendix. Figure 6 depicts the results. It lists the names of all 
relevant districts in the right-bottom listbox. Each district and 
its land use statistics are also highlighted in the map view. 
Figure 6. How the land uses along railway “Guang-Mei-Shan” 
change between 1987 and 1996 
Table 3 compares the amount of transmitted data when using 
different solutions for accomplishing this task. 
Server- 
side 
solution 
Client- 
side 
solution 
Layer- 
based 
solution 
The proposed 
solution 
Step 1 : 
Buffer 
7,892 
7,211 
7,211 (on 
server) 
1817 (on 
browser) 
Step2: 
Within 
92 
1,728 
1,728 (on 
browser) 
l,728(on 
browser) 
Stcp3: 
Stat. 
723 
3,946 
723 (on 
server) 
723 (on 
server) 
Total 
8,707 
12,885 
9,662 
4,268 
Table 3. Comparisons of the third case study (data amount is 
measured by Byte) 
4.3 Discussions 
The implementation of the above three case studies shows that 
the proposed solution is feasible and operable to enable load 
balancing spatial analysis in XML-based WebGIS. The 
comparison with server-side solution and browser-side solution 
also shows that load balancing solution can optimize the 
execution of spatial analysis, and therefore greatly ease the 
network transmission load between server and browser sides. 
With the proposed solution, high-performance spatial analysis 
can be easily provided in XML-based WebGIS. The comparison 
with our former solution (i.e., layer-based solution) shows that 
using spatial objects as a unit of organizing and transmitting 
spatial data can provide a better performance than using layers 
as a unit. It also reveals the importance of developing a more 
precise decision rule. To sum up, the proposed load balancing 
spatial analysis can enable users with high-performance spatial 
analysis in the Web environment.