Add a more pragmatic FAQ on network scenarios

[coffeemug]

Here are some questions that often come up:

• if one runs rethinkdb on multiple nodes, is fault tolerance built in?
• can i read/write to various nodes at the same time?
• is rethinkdb eventually-consistent?
• is multi-datacenter replication supported?
• can insertion/deletion/reads be done on separate nodes?
• can insertion/deletion/reads be done on separate nodes during network partitioning?
• what happens when a master dies? does it automatically pick a new master?
• what happens when the master comes back up?

We already have the answers in http://rethinkdb.com/docs/architecture/, but it takes some reading to get to them. We should have a more straight forward faq/answers to these questions so we don't have to answer them over and over again.

[coffeemug]

Also, this is loosely related to #10.