Location Redundancy

Location redundancy

One of the elements that we want to make sure is redundant is the location. Just as the name suggests, we want to create some sort of redundancy based off of location, so I'm going to have a whole other set of backup services located somewhere else.

Let's see what an example of this looks like. I'm in the Pacific Northwest, which is a great spot to host things. There are very few things that happen around here, there are very few natural disasters. So I'm going to turn up my services here. Another advantage to this is that it's very close to me, so I have very quick access to it. There's not a lot of geographical distance from where I live, and therefore I can manage this very quickly and easily.

But maybe I want to make sure my services outlast something that could happen here. Two of the most likely would be either an earthquake — because along the coast here there's a chance of a super powerful earthquake that could knock out a lot of the infrastructure — or some of the data centers that are over here are close to volcanoes, so if one of those volcanoes erupted then it could disrupt that communication.

So what I could do is put something as far a distance away from that as possible. Here, down in South Africa, is about as far as you could get away from where I live, so maybe I have some redundancy then that's located down there. There are even some mobile sites as well — there are mobile data centers that can move around, so that is a thing that exists as well.

Cloud providers make this affordable

One of the things that really can help facilitate this without having astronomical prices is by using cloud providers. Cloud providers make it really easy to spin up services in other locations. What this allows us to do is use cloud providers to create something that's over here in this area, and another area, and perhaps another area. We can have this geographical distribution without setting up whole data centers, without a huge investment into the infrastructure.

Amazon Web Services, or AWS, is one of those cloud providers that helps us really set up this redundancy without a lot of extra cost to it. It's relatively inexpensive to be able to turn up these new services across the world. They have something called regions and availability zones, and I thought we could take a look at what these regions and availability zones are, so we get an understanding of this geographical distribution that we can have here.

Regions and availability zones

Here we can actually see where their data centers are located. Right now North America is selected. We can see North America here — we've got one, two, three, four, five, six, seven, eight different locations here. So that's North America. I can select other regions as well. If I go to South America, you can see I've got a couple of options down there. If I choose Europe, I have many options in Europe. Middle East — I can see that I have a few options in the Middle East.

So these are regions. Each one of these dots is a region, and there's pretty far geographical separation from this, as you can see. But also, if I select some of these, there is Europe Stockholm, and we can see that there are three availability zones. What that really means is there are three data centers.

What these data centers essentially are is separate buildings. They're located somewhat near each other, but they're still separate buildings, so they have whole different infrastructures available. So it guards against some things that could happen, but not everything, right? It's not geographically separated by countries, it's just geographically separated within an area, by building — it's separated in buildings.

Now, the advantage is that there are really fast speeds between these buildings, and it's really easy to do failovers, and this really sets it up for having two live sites right next to each other. The further you get away, the chances are that there could be a little bit more delay, which can cause some problems on some applications if they're not built right, or if they're built in a certain way.

So what would be great here is to set up two live production sites that are side by side, that are running, that are doing some sort of load balancing between two different availability zones, but then have maybe a hot or warm site somewhere else. Now, what do all those terms mean? It doesn't really matter at this point in time. But the point with all of this is that having some sort of geographical separation between redundant sites is actually a great thing to have.