Test Setup:
- Four Fedora instances running individually in four AWS EC2 instances and are configured in replication mode.
- A EC2 load balancer is configured on top of these Fedora instances.
Procedure:
- Started a single instance of Fedora in replication cluster mode.
- Loaded 6000 objects with unique datastreams.
- Used JMeter to perform response time test directly on a single fedora instance. (1 nodes)
- Added two new Fedora instances to the cluster.
- Add these three Fedora instances into EC2 load balancer.
- Used JMeter to perform response time test via load balancer on the cluster. (3 nodes)
- Add one more Fedora instances into EC2 load balance.
- Used JMeter to perform response time test via load balancer on the cluster. (4 nodes)
- Test cases
- 1000, 3000, 6000 and 10000 HTTP requests for objects.
Results:
Response time of 1000 Requests in 60 seconds
| Single Fedora Instance | 3 Node Cluster | 4 Node Cluster |
|---|---|---|
 |  |
|
Response time of 3000 Requests in 60 seconds
| Single Fedora Instance | 3 Node Cluster | 4 Node Cluster |
|---|---|---|
 |  |  |
Response time of 6000 Requests in 60 seconds
| Single Fedora Instance | 3 Node Cluster | 4 Node Cluster |
|---|---|---|
 |  |  |
Response time of 10000 Requests in 60 seconds
| Single Fedora Instance | 3 Node Cluster | 4 Node Cluster |
|---|---|---|
 |  |  |
| Test Setup | Num. Requests | Test Duration (Seconds) | Average Response Time (ms) | Max Response Time (ms) |
|---|---|---|---|---|
| Individual | 1000 | 60 | 38 | 4203 |
| Cluster (3) | 1000 | 60 | 38 | 4214 |
| Cluster (4) | 1000 | 60 | 38 | 4246 |
| Individual | 3000 | 60 | 21 | 80 |
| Cluster (3) | 3000 | 60 | 27 | 4237 |
| Cluster (4) | 3000 | 60 | 26 | 4266 |
| Individual | 6000 | 60 | 21 | 79 |
| Cluster (3) | 6000 | 60 | 24 | 4310 |
| Cluster (4) | 6000 | 60 | 23 | 4212 |
| Individual | 10000 | 60 | 21 | 584 |
| Cluster (3) | 10000 | 60 | 23 | 4213 |
| Cluster (4) | 10000 | 60 | 21 | 200 |
