Fault Tolerance in Grid Computing

1
Fault Tolerance in Grid Computing

• Presented By
• Jyothi Reddy

2
What is Grid Computing

• Its a special type of parallel computing.
• Different from Supercomputers, but can be called
  virtual SuperComputers.

3
(No Transcript)
4
Applications of Grid Computing

• Enabling Grids for E-SciencE project CERN
• SETI_at_home
• NASA Advanced Supercomputing Facility (NAS).
• World Community Grid Project

5
Recent Work in Grid Computing

• Fault Tolerance Scheme for Hierarchical Dynamic
  Schedulers in Grids.( IEEE Computer Society 2008)

6
Scheduling Methods

• Static Scheduling
• Dynamic Scheduling

Mapping done depending on resource environment
situation during the scheduling time.
Mapping done depending on resource environment
situation when the task is ready to be executed.
7
Fault Detection and Recovery
1
3
5
6
8
9
11
8
Fault Detection

• Fault Detection is done by heartbeat messages
  sent from node to immediate descendant
• Latency is detected observing Round Trip Time

9
Recovery

• Node n forms a cluster with n as cluster head.
  n,k,l Where k, l are descendants.
• Where k is called the associate cluster head.

10
My Research

• Trying to improve the existing Algorithm (or)
• Hoping to use a genetic algorithm(GA) to the
  nodes for effective scheduling

11
Thank You

• Citations
• A Fault Tolerance Scheme for Hierarchical Dynamic
  Schedulers in Grids
• IEEE Computer Society, Nitin B. Gorde, Sanjeev K
  Aggarwal, Dept of Computer Sciences, IIT Kanpur
• Wikipedia.org, Fault Tolerance in Computers.
• http//en.wikipedia.org/wiki/Fault-tolerant_system
  
• 3. Grid computing, Wikipedia.org
• http//en.wikipedia.org/wiki/Grid_Computing