Experiments on R-trees |
|
|
- Split time increases with page size
|
|
- Decreases with increasing
|
|
|
- Becomes very high with large m due to lots of re-inserts
|
|
|
- Relatively insensitive to splitting algorithms
|
|
- Less random I/O cost but more CPU cost with larger page size
|
|
- Slight effect of : increases with
|
|
- Size of index decreases with increasing
|