Proceedings PaperImpact of dynamic design on data retrieval speed
|Format||Member Price||Non-Member Price|
|GOOD NEWS! Your organization subscribes to the SPIE Digital Library. You may be able to download this paper for free.||Check Access|
The design of the link between data and its data description defines the flexibility and application. A static configuration refers to a situation where metadata are externally defined. Conversely, in dynamic cases the data descriptions are no longer frozen, but are explicitly formalized and stored at a certain hierarchy level. We used two levels in order to test dynamic design; catalogs as dynamic lists and images as dynamic items. We then considered three kinds of objects, always containing a static data part and eventually a dynamic data part. Using the object oriented database Objectivity, we measured the retrieval speed for several configurations. When selection criteria apply to the static part of items, the retrieval speed is independent of the data kind extracted (static or dynamic). However, when the selection criteria apply on dynamic parts, the speed is strongly decreased. This clearly shows the strength of static implementation of informations whenever it is possible in order to guarantee a fast data access. It also point out a serious limitation to data mining, where a priori knowledge is in general not available. Fortunately, a dynamic implementation at a level of lists, could resolve the problem. The most elegant way, in the VO context, would be the usage of VOTable as data access layer interface.