In this thesis I propose an automated data replication system that copies the frequently used data from less reliable servers to the most reliable ones in the distributed system. The replication decision is being taken by an algorithm that ranks the servers in the system from unreliable to reliable. This rank is being calculated based on their crash history logs of the servers, reliability of the hardware, size of the data, and the distributed model used in the system.
The outline of the paper is mostly done (can be changed if writer has better suggestions)
Problem
• High availability
• Fault Tolerance
o Outages
Planned (need for maintenance of the system)
Unplanned (caused by faults)
o Downtime
Direct cost of DT
Indirect cost of DT
• Redundancy
• Data Replication
Abilities
• Ability to run in a heterogeneous environment.
• Ability to provide optimum data availability without creating data redundancy
• Ability to decrease overhead on updates of the replicated data
How to do it
• Check hardware
o Oldest one
o Fails a lot
• The role of the server
o Main data or important data
What to do
• Rank the servers
• Replicate the data to the most reliable one based on the rank
Challenges
• Consistency
Research for availability is done
fault tolerance is mostly done. Other parts has not done yet.
8-9 pages is already written. This written part includes Availability and parts of fault tolerance.
I need professional help to get this thesis done by may so I can defend it by the end of the semester (early-mid May)
Use the order calculator below and get started! Contact our live support team for any assistance or inquiry.
[order_calculator]