High availability depends on having the right level of redundancy, measured using N+1, N+2, or N+N models that define how many spare systems sit alongside active infrastructure. The appropriate redundancy tier balances the cost of additional hardware against the potential revenue impact of unplanned downtime.
Redundancy for High Availability
If we want to achieve high availability, then we're going to need to have some redundancy within our infrastructure. But how much redundancy do we need?
Sure, we want redundancy, but the question is how much redundancy do we want? If we need one server, do we set up two servers? Do we set up three servers? Do we set up four servers? How many extra servers do we want to set up?
One of the ways that we measure this is with N+1, or two, or three. What we say is we want to achieve a certain level of having spares available. So N+1 means whatever number we have, we want one more than that.
Let's take a look at a non-technology example of N+1. Let's take a look at a car. What does a car need to go down the road? Well, a car has four tires. And so if we have four tires and it's required to have four tires to go down the road, then we have one spare. N+1 means that we have four - that's the N part of that equation - and the plus one is that extra spare tire to replace one of those tires if it goes flat.
So in this example right here, maybe we need two servers that are up and running. Well, if we need two servers up and running, we're going to have a spare server. So if any one of those goes down, we can move the load and the processing of that server to that third server. Same thing here: we maybe need one database server. And if we just need one database server, then we have one on standby. So this is how many we need, which is one, plus one extra.
So you always have one extra with N+1. N+2 means we have two extra spares. You can see that N+3 is three extra spares, and N+N means that we have a duplicate of everything. So if we have five servers, then we have an extra five servers on standby.
Now, in my example of the car, the plus one is something that's not really operational until you put it on. As an example of that, I have four tires on a car, one of them goes bad, I take this spare that was not being used out of the trunk and then I put it onto the car. So that is the plus one right there.
But we could be utilizing that extra spare, that redundancy there. And what does that look like? Well, here we have the need for a server. Let's say we determined we need the use of just one server, but we're practicing N+1, so we add a second server. It doesn't have to be turned off. It could actually be functioning, it could actually be working. And what we do is we just split whatever load we have: half goes to one server and the other half goes to the other server.
What's great about this is that we know that both servers are functioning. And when one goes offline, we just move the 50% of the load that was going to the second server and we move it onto the first server. So this is still an example of N+1.
So which one are we actually going to choose? Well, a big part of this is going to come down to cost. Something that has just a little redundancy to it is going to cost much less than something that is fully duplicate. You have to pay for twice as much stuff with this N+N. So a lot of it is just going to be the cost and how much you are willing to spend on it.
The other side of this equation is how much impact is it going to have if you don't have this redundancy? If you have downtime, how much does it impact the business? So if you're going to have a huge amount of revenue loss because you're going to have downtime in IT, then you're going to pay for a lot more redundancy on your network.
TechKnowSurge builds IT and cybersecurity professionals through hands-on, concept-first training built around real understanding — not memorization. Free interactive tools, structured programs, and 25+ years of real-world experience, all in one place.
Explore free tools and programs →