Test setup

The performance cluster configuration contains one machine and one instance for each interviewing role: Accessory, Web, Interview, Database, and FMRoot. Three additional machines are required to run the load tool. Two very similar clusters were used for the testing. See Cluster configuration for details.

The projects in the following table were used in load tests at loads of 1, 10, 20, 100, 200, and 400 simultaneous interviews to check the resource usage for different numbers of levels and interview load.

#	Project	Levels	Categories	Designed questions	Variables in database	Actual asked questions	MDD file size
1	HData3LevelUnexpanded.mdd	3	1003040	4	39	65	309 KB
2	HData3LevelExpanded.mdd	3	1003040	4	123,130	65	309 KB
3	HData5LevelUnexpanded.mdd	5	1010101010	6	45	63	310 KB
4	HData5LevelExpanded.mdd	5	1010101010	6	111,140	63	310 KB

The projects in the following table were used with a load of 100 simultaneous interviews to check the impact of different numbers of VariableInstances.

#	Project	Levels	Categories	Designed questions	Variables in database	Actual asked questions	MDD file size
5	HD3U3H.mdd	3	1056	4	39	65	270 KB
6	HD3E3H.mdd	3	1056	4	390	65	270 KB
7	HD3U2T.mdd	3	25810	4	39	65	275 KB
8	HD3E2T.mdd	3	25810	4	2,255	65	275 KB
9	HD3U15T.mdd	3	501520	4	39	65	284 KB
10	HD3E15T.mdd	3	501520	4	15,830	65	284 KB
11	HD3U50T.mdd	3	802030	4	39	65	295 KB
12	HD3E50T.mdd	3	802030	4	49,710	65	295 KB
13	HD3U1M.mdd	3	10030040	4	39	65	303 KB
14	HD3E1M.mdd	3	10030040	4	123,130	65	303 KB

#	Project	Test	Levels	MDD file size
15	HDataLargeUnbound2.mdd	Resource usage for reading and writing loop questions during start/restart and when displaying the question	1	312 KB
16	HDataLargeUnbound3.mdd	Resource usage for reading, writing, and setting styles on loop questions and displaying the questions	1	313 KB

The diary study example, Diary_hdata.mdd, and the HdataUnbound_FunctionalityTest.mdd used to test functionality are also available in the DDL.