2. How do highly-distributed databases create problems with data consistency?
Increasing the number of servers involved increases the likelihood of a failure, both at the server and with the connecting networks. It also increases delays in synchronizing data to ensure all machines are updated. If consistency requires all servers to have the exact same data, a large number of copies distributed across thousands of computers will result in processing delays while waiting for all machines to respond to update messages. It also reduces the benefits of distributing the processing and data storage. By working independently, portions of the data can be updated on different machines at the same time. Forcing transactions/updates to be universal slows all updates down to the slowest process.