The distributed system can achieve:
- high availability
- high performance
The biggest fallacies in our abstracted distributed system are:
- The network is reliable.
- Latency is zero.
- Bandwidth is infinite.
- The network is secure.
- Topology doesn’t change.
- There is one administrator.
- Transport cost is zero.
- The network is homogeneous.
The Consistence Model
Once multiple copies of the data are saved in the distributed system, — a must for the high availability, — you have to deal with the consistence problem.
One approach is Single-copy Consistency Model, at all the time, only one write is served. Alternatively, you may apply a Weak Consistency Model to trade for Availability and Partition tolerance.
- Distributed systems theory for the distributed systems engineer
- Distributed Systems for Fun and Profit
- Petr Bailis’ blog
- Life Beyond Transactions
- Google I/O 2009 - Transactions Across Datacenters