2004 
rver 
blic 
OXY 
e in 
lave 
atial 
s of 
data 
atial 
d. If 
ime, 
atial 
will 
yer, 
obal 
yer, 
che, 
ding 
ined 
atial 
slots 
, We 
yer. 
eda 
the 
the 
will 
ules 
ayer 
Fr a 
of 
way 
; we 
very 
slots 
's in 
nnot 
( 3 
;, its 
; the 
of a 
ities 
rger 
into 
line. 
t the 
  
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B-YF. Istanbul 2004 
same time as updating spatial entities in a layer in spatial 
database. Offline refreshment means updating invalid spatial 
data of a layer in cache if their versions are invalid when 
accessing to them. 
Information in spatial cache includes /ayerVersion, slotVersion, 
entityVersion and the corresponding relations between all 
entities and slots of a layer. When accessing to entities in a 
layer, we can determine whether spatial data in cache is valid or 
not through comparing version information in cache and spatial 
database. 
(1) If /ayerVersion of a layer in cache is smaller than that in 
spatial database, it means that all entities of the layer are 
changed, and the whole layer should be refreshed in cache. 
Otherwise, 
(2) If slotVersion of a layer in cache is smaller than that in 
spatial database, it means that all slots of the layer are 
rearranged, and the corresponding relations between all entities 
and slots of the layer should be refreshed in cache. Otherwise, 
(3) If entityVersion of a layer in cache is smaller than that in 
spatial database, it means that some entities of the layer are 
changed, and those entities should be refreshed in cache. 
(4) If above three conditions are all equal to those in spatial 
database, it means spatial data of a layer in cache is same as 
that in spatial database, and no refreshment is required. 
When accessing to spatial data, if we can pre-load some spatial 
data into cache, it will make spatial cache more effective. But 
how to predict what kind of spatial data will be accessed to? 
There are two rules of accessing to memory: there is every 
probability of accessing to just being accessed memory, and 
there is every probability of accessing to neighbors of just being 
accessed memory. There are same rules in accessing to spatial 
data: there is every probability of accessing to just being 
accessed spatial data, and there is every probability of accessing 
to neighbors of just being accessed spatial entities. According to 
those rules, we can arrange slots by spatial scope, and a slot is 
an accessing unit of spatial data. When the network is idle, we 
can pre-load some neighbors of spatial data in cache from 
spatial database. 
3. CONCLUSIONS 
Geo-Union has finished a preliminary component-based model 
for distributed WebGIS, and has got into use in many fields 
with sound effects. Although spatial cache and other techniques 
are adopted in Geo-Union to improve its performance, a lot of 
works still wait us ahead to make Geo-Union more practicable 
and effective: 
(1) Dynamic load balancing policy. In Geo-Union, most works 
are completed at client side, and server is only responsible for 
data access and simple data query. Therefore it is not well 
balanced between client and server. Especially there will exists 
a lot of transmission for massive data between client and server. 
Although, spatial index and spatial cache techniques can 
improve the performance to a certain extent, we still want to 
take full advantage of the computing capability of GIS server, 
so as to lighten load at client side and decrease the transmission 
203 
of redundant data in network. This needs a more reasonable 
component design for the system. 
(2) System concurrency. WebGIS is open to millions of users. 
How to ensure the correctness, validity, stability and scalability 
of Geo-Union to satisfy users’ requests is another key problem 
for practicable WebGIS. 
4. ACKNOWLEDGEMENT 
Supported by the National Grand Fundamental Research 973 
Program of China under Grant No.2002CB312000; the National 
Research Foundation for the Doctoral Program of Higher 
Education of China under Grant No. 20020001015; the National 
Science Foundation of China under Grant No.60203002; the 
National High Technology Development 863 Program under 
Grant No. 2002AA 135330 and No. 2002AA 134030: the Beijing 
Science Foundation under Grant No.4012007. 
5. REFERENCES 
Li Bin, 2001. A Component Perspective on Geographic 
Information Services. — Cartography and Geographic 
Information Science, 27(1), pp.75-86. 
Dept. of Computer Science and Technology, Peking University, 
2002. Operation and Component Guide for Geo-Union 
Enterprise (in Chinese). Technology Material, 
http://gis.pku.edu.cn. (accessed 20 Oct. 2002).