2 hours 38 minutes

Video Description

The next topic under spares and redundancy is clustering and web farms. In the previous section we covered RAID and how it provides redundancy for disk drives, but what about doing the same for servers? This is addressed with redundant servers or clustering: multiple servers acting as a single logical unit. An example of this in action is Amazon. It may appear that you're only connected to a single server when you're shopping from the comfort of home, but behind the scenes, there is a multitude of servers handling your requests! Clustering is primarily for fault tolerance, but another implementation also provides load balancing for performance optimization. We point out that clustering and load balancing are not the same and that various implementation are available to support redundancy, load balancing, or both.

Video Transcription

Okay, So Ray is gonna protect us against hardware. Hard drive failure. But what about if other hardware on their server fails? What if the processor overheat? Would if we have problems with the power supply or some other issues? Well, we want redundant servers.
So, uh, you know, one thing to think about for redundancy is
you know, if if my domain controller fails, then we can access the domain. Well, have a second domain controller so we can have redundant servers. And service is on the network for that specific purpose. We can also
bring in what we call a cluster. We can even use clustering to provide his redundancy of our servers.
So here we truly have multiple servers acting as a single logical unit. Whereas in the slide before we were talking about redundant servers, we very distinctly have two different servers here. We may have five notes and a cluster acting as a single entity. And appearing is a single entity to users,
you know? Think. For instance, when you go to Amazon
www dot amazon dot com Now, how many servers or dare waiting to respond to your request?
Eight zillion. Right. But it feels like I'm just connecting to one system that's a Web cluster. Sometimes you'll hear the term server farm, and that's very appropriate as well. Now clustering is primarily designed for fault tolerance. That's its job. If one server in the note fails,
the other notes pick up the work.
Clustering very frequently also does load balancing, but the two are not the same. Clustering multiple nodes. Acting is a single logical entity. Load balancing adds thehe bility to distribute the work across those servers. Not all clusters do that.
You can have an active passive cluster
where you've got the primary who responds toe all the request is active. And then the second servers just there, waiting for the heartbeat from the server,
which is basically just a little message that says, Still here.
Still here.
And as long as it continues to hear the heartbeat, the passes, the passive server states path. But should the heartbeat fail for certain amount of time, the passage the passive server would kick in and begin processing requests.
The reason I tried to stress that as many people think all clusters, dude load balancing and load balancing is always on clusters,
and that's not the truth of the matter. So the clusters just multiple servers. Acting is a single logical entity

Up Next

Security Operations

They are responsible for knowing where a network's possible vulnerabilities are and providing mitigation strategies to combat them. An effective Cyber Security Operations Manager will have experience in a technical security role including ...

Instructed By

Instructor Profile Image
Kelly Handerhan
Senior Instructor