Baseline report for DSpace 6.3
During this test, the number of concurrent threads started at one, and was doubled after 12 minutes, doubling every 2 minutes after that, peaking at 32 threads, then ramped down to a single thread for a couple minutes to end the test.
For detailed information about the user activities modeled in this test, see Sessions, transactions, and requests.
Single-threaded results
During the first phase of the test, a single thread was used to make all requests, so the server only had to handle one request at a time. This put an artificially low load on the server, while providing an average “best case” roundtrip time number per request.
IMPORTANT: single-threaded != single-user
The seconds/session numbers reported below were computed by adding all delays for loading each resource for each page. A real browser would spend less time waiting for all resources to load due to parallel loading of resources. So, while helpful in understanding the overall performance picture, the seconds/session numbers do not accurately reflect expected page wait times.
Session | Pages/session | Bytes/session | Seconds/session | Transaction details |
---|---|---|---|---|
Big | 7 | 891KB | 4.3 | See details |
Browse | 6 | 1.0MB | 21.2 | See details |
Search | 5 | 925KB | 1.5 | See details |
Tiny | 7 | 882KB | 2.7 | See details |
Test results over time
The following graphs show how certain measurements changed over the lifetime of the test.
Homepage load time
This shows how the average time to load all home page resources, back-to-back, changed over the life of the test, depending on how many threads were active.