Hardware Redundancy

Redundancy in a server

This is a server, and really we can turn any computer into a server. That is, we can turn a computer into offering some sort of service. But this is a specific physical server. It's designed for enterprise level servers and services. Why is that the case? One of the reasons is because it has a lot of redundancy on it.

Right off the bat, what we can see here is that there's a lot of these similar things on the front, and what these are are drives. I have two of these drives mirrored, and one of them is just sitting in there not mirrored. What that means is that two of these drives are actually redundant — that one drive is the same as the other drive, that's the exact same thing. So this is one example on this server of how we can create redundancy within those drives, within the storage on the server.

Next, we're going to actually open this up. I'm going to put this drive back in here, and we're going to pull this up and take a look on the inside. On the inside, what we can see is that there's a lot of duplicate parts on here. One of those things are fans. There's lots of fans on here, and they just pull out like that, so we could replace those fans if we wanted to. In fact, most of this stuff is modular.

In the back here we have only one NIC card, and I'm not going to pull that right now, but this is an example of where, if I wanted to have an enterprise level, I would need two network interface cards back here. This is just for my home server.

Then we see two processors. That's because we could utilize two processors, so that's not necessarily redundant. Same thing with the RAM that we've got here — it can actually handle a lot of RAM. But there are components in here like the power supply. So here is the power supply. Let me slip this out right here. We can see this power supply, and if you notice, there are two of them. So we could have a power supply that could fail on this. So we see one of the advantages of buying an actual physical server is that we can create this redundancy within here.

There's one other thing that I want to point out here, and that is the PERC card. These drives right here, as I mentioned, are capable of handling — I think it's eight drives. It's capable of eight drives right there. But there's a card that manages all of that. I think it's this one right here. It's called the PERC on this one, and really it manages the RAID configuration. We'll get into what RAID is, but it is what allows us to have redundancy within these drives, and there's different configurations that we can do. This is a hardware RAID on here.

So here's just some examples on a server, some redundant parts of the server to make sure it stays up and running.

Redundancy in a switch

This right here is an enterprise level switch. You can see that it's 48 ports, plus we have the capability of sliding some extra capabilities into here. As we see it right now, this doesn't really necessarily have redundancy just by first glance, but it could be redundant if we just connect more Ethernet connectors. So if I'm connecting two switches, I can link the two with more Ethernet, so if one of those links goes down then it's still up and running.

But if we turn it to the back, what we can actually see is there is an amount of duplicate with this in the form of power supplies. So we see two power supplies right here that I can pull out, and this is redundant. We could have one of these power supplies go out and this would still be up and running. We can also see two fans here, so I could pull the fans out. We see two of them right here. So that's another example — there's some redundancy within the system.

StackWise

But there's another thing that I want to show you. There's two switches, and these are special types of switches. They're Cisco switches and can do something called StackWise, which creates some redundancy amongst two switches.

On the back here, we have these StackWise ports, and this is a StackWise cable right here. What I can actually do is plug in these StackWise cables like this, and I can link quite a few of these switches together and stack them all up, and essentially they all become one switch. I can even loop the top one back down to this bottom one, because there's two of these ports for each one.

What this does is it interlinks the backplane of this. It's a much faster connection than Ethernet ever would be. And then they all act as one switch, but even if one of these switches failed within this one massive switch, then the rest would still be up and running. So what it allows us to do is have some redundancy amongst these switches.

Redundancy with firewalls

Now, these are enterprise level firewalls. If you take a look at this, at first glance they don't really have redundancy involved with them, but there's two of them, and so it creates a redundancy with these two pieces of hardware.

What you would do is you would program one of these ports to be a heartbeat, and the heartbeat you would connect together. You'd have to use a crossover cable, but you would connect it together. Then that heartbeat just constantly communicates with each other, and then you would configure it to be a failover of each other. So then what they would do is they would communicate back and forth — hey, are you still up? Hey, are you still up? Hey, are you still up? Then if the other one doesn't respond within a certain amount of time, then it does this failover and the other one becomes the live switch. So we can have redundancy amongst these ASA firewalls through that heartbeat and through a failover process.

So those are some examples of how we would set up equipment to fail over. Some of it is components within the equipment that is set up for redundancy. Some of it you have to set up specifically to be redundant amongst this equipment.