SuperMap SDX + engine optimization

This section mainly introduces the optimization strategy of SuperMap SDX +. You can choose your own optimization method according to your needs.

Hardware platform still plays an important role in the performance of GIS system, but the more expensive the hardware is, the more suitable it is. We must make a reasonable plan according to the needs of the system and choose a cost-effective hardware platform suitable for our own GIS system.

The following are the options you can refer to when selecting a hardware platform:

File caching is an intelligent distributed storage solution provided by SuperMap SDX + to balance network and server loads and improve the overall performance of applications. After the file cache function is enabled, when the Application accesses the data stored in the Spatial Data library or the network server, it will first check whether there is a Newest Version of the corresponding data in the local cache library. If there is no corresponding Cache Data or the Cache Data is not the Newest Version, the data is read from the server side and the local Cache Data is updated, so that the local Cache Data can be read directly during the next access; If the local cache already has the Newest Version of the corresponding data, it is not necessary to request the data from the server through the network, but to read the local Cache Data directly to complete the display or analysis function. With this Solution, the database server load and network load can be significantly reduced, thereby significantly improving the overall performance of the Application.

The following is an example of the difference between using cached mode and not using cached mode Display Speed. Test data: PolylineDataset (Records: 690,000) of 1:500 topographic map of a city; server: common PC compatible machine, CPU (P4 2.4), memory (DDR 1024M), hard disk (7200 rpm); client and server are the same machine.

The following tests are Data Full View, run three times and average:

表1：缓存与非缓存模式Full View时间对比
Pattern	Item	Time (seconds)
Uncached mode	The first time	9.50
	The second time	9.41
	The third time	9.32
	Average value	9.41
Cache mode	The first time	3.56
	The second time	3.40
	The third time	3.50
	Average value	3.49

As can be seen from the above table, the Display Speed of cached mode is much faster than that of non-cached mode. Therefore, if the data does not change frequently, please select to use file cache mode.

The value of Layer's MinVisibleGeometry Size property determines the minimum visible size, in pixels, of the Geometry. The default value is 4. When this property is set, the object is not displayed if its display size is smaller than this value.

For Dataset with a large amount of data (more than 50000 records), such as lines, surfaces and networks, setting the appropriate MinVisibleGeometry Size attribute value for Layer can achieve good optimization results.

Test data: PolylineDataset (Records: 690,000) of 1:500 topographic map of a city

Server: PC compatible, CPU: P4 2.4 Memory: DDR 1024M Hard disk: 7200 RPM

Client: Normal PC compatible, CPU: PIII 667 Memory: 256 Hard disk: 5400 RPM

The following tests are Data Full View, run three times and average:

表2：不同的MinVisibleGeometrySize下Full View的时间
	0	1	2	3	4	5
The first time	19.75 seconds	15.60 seconds	14.62 seconds	13.81 seconds	13.65 seconds	13.51 seconds
The second time	19.59 seconds	15.28 seconds	14.84 seconds	13.90 seconds	13.91 seconds	13.66 seconds
The third time	19.31s	15.41 seconds	14.32 seconds	13.91 seconds	13.85 seconds	13.44 seconds
Average value	19.55 seconds	15.43 seconds	15.43 seconds	13.87 seconds	13.80 seconds	13.64 seconds

Through the above comparison, the following conclusions are drawn:

For your system data, you can also find the best point of entry for performance and Display Effects through comparative testing. The recommended range of MinVisibleGeometry Size of Layer is 0 ~ 5.

SuperMap SDX + provides four Spatial Indexes to speed up data access, including:

SuperMap SDX + automatically builds a Spatial Index for the Data when it is imported into the database (in Newest Version, the default index type is a multi-level grid index), and you can build other indexes as needed.

With the change of data (addition, deletion and modification), incomplete Spatial Index will appear. When the amount of modified data is very large, the access speed of data will be affected. At this time, the index needs to be updated or rebuilt. IsSpatialIndexDirty determines whether the Spatial Index in the Vector Dataset needs to be rebuilt. If the value of IsSpatialIndexDirty is TRUE, the index needs to be rebuilt; If the IsSpatial IndexDirty value is FALSE, there is no need to rebuild the index.

In the process of using SuperMap for Secondary Development, if the value of a field needs to be used frequently for positioning, query and other operations, if an index is established for such fields, the efficiency of program operation can be greatly improved.

SuperMap's Image Pyramid technology can be easily competent for the organization and management of massive Image Data. Using SuperMap Image Pyramid technology, you can achieve smooth roaming of massive Image Data that is independent of the amount of data and the display area.

For larger Image Data, the Create Image Pyramid can significantly increase the Display Speed of the Image Data. Examples are as follows, data file: Beijing area (local) 547 km2 remote sensing Satellite Image (10 maps in total); Pixel Format: 24-bit true color; total number of pixels: 2606464372; Source Data size: 7457.15438M, the following table shows the comparison of the full display and the enlarged Display Time of the Create Image Pyramid and the Create Image Pyramid.

表3：Full View和放大Display Time对比
	Not Create Image Pyramid.	Create Image Pyramid.
Full View	14 minutes and 28 seconds	0.5 second
Magnify 2 times	6 minutes and 35 seconds	0.5 second
Magnify 4x	4 minutes and 10 seconds	0.5 second
Magnify 8x	1 minute 57 seconds	0.5 second
Magnify 16x	1 minute and 02 seconds	0.5 second
Magnify 32 times	24 seconds	0.5 second
Magnify 64 times	8.06 seconds	0.5 second

It can be seen from the above table that for nearly 7.5G Image Data, SuperMap can ensure that the Display Time of browsing in any area and any range is controlled within 1 second, which provides a reassuring Solution for the organization, management and use of massive Image Data.

Select the appropriate hardware platform
SuperMap SDX + data engine needs to access a large amount of Spatial Data. In the process of data access, the performance of memory access is much higher than that of disk access. Therefore, both database system and SuperMap SDX + make full use of memory to operate and process as much as possible, and cache some data in memory. Therefore, no matter for relational database system (DBMS) or SuperMap SDX +, Increasing memory capacity significantly improves performance.

In the process of data access, the hard disk needs to be accessed frequently, so the performance of the hard disk has a direct impact on the performance of SuperMap SDX +. The data is placed on the hard disk with high rotation speed, which can obtain higher data access speed.
Network Data Engine and Spatial Database Engine
in SuperMap SDX + are real-time Spatial Data access technologies, and the network speed has an important impact on the performance of data transmission. For example, the performance of a GIS Application system based on Spatial Data Library technology running on a 100 M network is significantly better than that on a 10 M network. The greater the number of concurrent users, the higher the network bandwidth requirements. In addition, in a local area network environment with a large number of concurrent users, the backplane bandwidth of the network switch or hub is also very important. For example, in a 100 megabyte local area network environment, a backplane bandwidth switch or hub of 1000 megabytes or higher can be considered.
- Memory
- Hard drive
- Network
Choose to use the file cache feature
Set the appropriate MinVisibleGeometry Size property value for Layer
- Layer, the lower the level of detail displayed;
- Layer, the faster the display speed;
Create a Spatial Index
1. Index, which supports File Database and Database Data source. The spatial retrieval efficiency of R-tree Index is very high, and its accuracy is also high, but the update of R-tree Index is more complex, so it is more suitable for static data.
2. Index, supporting File Database and Database Data source. It is suitable for high concurrent editing of small data volume.
3. Multi-level grid index, only applicable to database Datasource. The data is organized and managed by dividing a multi-layer grid. When the Dataset is being browsed, the speed of the multi-level grid Index Type is relatively fast, and the multi-level grid index has good update and concurrency capabilities, and the precision and accuracy of spatial retrieval are high;
4. Maps heet Index, only for database Datasource. The spatial objects are classified or displayed according to the fields or Range Settings. It is mainly used to process the framing data and is suitable for massive data. The combination of Mapsheet Index and file cache greatly improves the speed of data display and query. It is more suitable for Data Browse than Edit Data.
Create a Field Index
About Optimization of Image Data