In the previous section, we talked about the single point of failure and why having multiple instances of the application server is a good thing. We can extend this concept further; what if we have all of our servers in one location and due to some major malfunction or outage, all of our servers go down? We will not be able to serve any user traffic.

We can see that having our servers in one location also creates a single point of failure. The solution for this would be to have application server instances available in multiple locations. Then the next question would be: how do we decide on the locations to deploy our servers? Should we deploy the servers to multiple locations within a single country or should we deploy them to multiple countries? We can rephrase the question using cloud computing terminology as follows.

We need to decide whether we want to deploy our servers to multiple regions or multiple zones, or perhaps a combination of both.

One important point to note is that deploying to multiple zones may lead to network delay and we may want to deploy to multiple regions first. However, before we deploy to multiple regions and zones, we need to make sure of two facts:

• Our website has heavy traffic that our single server setup is no longer able to handle
• We have a significant chunk of users from another country, and it might be a good idea to deploy servers in a zone near them

Once we have given consideration to these factors and decided to deploy to additional zones and regions, our blogging system as a whole might look something like this: