L-Store Upload Structured Testing: Difference between revisions
Jump to navigation
Jump to search
| Line 27: | Line 27: | ||
Please watch this page as we post results from our runs. | Please watch this page as we post results from our runs. | ||
= Results for Small Uploads with Three different thread sizes and Small Number of Files = | |||
{| class="wikitable" style="text-align:center" | {| class="wikitable" style="text-align:center" | ||
* Umich-Umich (MBytes/second) | * Umich-Umich (MBytes/second) | ||
|- | |- | ||
!bgcolor=red | <font color=white> | !bgcolor=red | <font color=white> Type of Test || bgcolor=blue | <font color=white> Number of Threads || bgcolor=grey | <font color=white> Average File Size || bgcolor=grey | <font color=white> Average Latency || bgcolor=grey | <font color=white> Median Latency || bgcolor=grey | <font color=white> Maximum Latency || bgcolor=grey | <font color=white> Std Deviation on Latency | ||
|} | |} | ||
Revision as of 15:31, 15 June 2006
Objective
This is the first set of tests for Benchmarking L-Store system performance. We are setting up Upload testing to begin with.
Parameters
1. File Size:
* Small: 1KB file will be used * Medium: 0.5MB file will be used * Large: 260MB file will be used
2. Number of Files
* Small: Essentially we will use the number of uploads set up as a surrogate for number of files. 50 * Medium: 200 * Large: 1000
3. Number of threads
* Small: 1 * Medium: 20 * Large: 60
- Experimental Design: Full Factorial (3*3*3=27 test cases). Each test case would be run 30 times to collect statistically significant data.
Please watch this page as we post results from our runs.
Results for Small Uploads with Three different thread sizes and Small Number of Files
- Umich-Umich (MBytes/second)
| Type of Test | Number of Threads | Average File Size | Average Latency | Median Latency | Maximum Latency | Std Deviation on Latency |
|---|