From: Fault tolerance in computational grids: perspectives, challenges, and issues
S. no. | Evaluation criteria | Recommended | In between | Not recommended |
---|---|---|---|---|
1 | Time to detect errors | Early | – | Late |
2 | Failure probability | Low | Medium | High |
3 | Node selection for job execution | Intelligent | Random | Unintelligent |
4 | Failure detection | Proactive | Reactive | – |
5 | Fault detection layers | All layers | Few layers | No layer |
6 | Recovery time of failed node | Low | Medium | High |
7 | Response time after failure | Early | – | Late |
8 | Resource utilization | Increased | – | Decreased |
9 | Recovery technique | Workflow, task level | – | – |
10 | Job success ratio | Increased | Moderate | Decreased |
11 | Overall throughput | Increased | Moderate | Decreased |
12 | Overall ATAT | Reduced | – | Magnified |
13 | Errors detected | Large | Medium | Low |
14 | Overall AWT | Low | Medium | High |
15 | Transmission delay | Reduced | – | Magnified |
16 | Implementation | Easy | – | Difficult |
17 | Adaptability | Yes | – | No |
18 | Fault detection | Dynamic | – | Static |
19 | Task level FT | Checkpoint | Replication | Alternate resource |
20 | MTTF | Increased | – | Decreased |
21 | MTTR | Decreased | – | Increased |
22 | MTTD | Decreased | – | Increased |