Funny ways to explain things… Part II (VMware)

Bookmark and Share

My first blog post on how to explain technical things to non-technical people I talked about how I explain storage concepts surrounding EqualLogic and VMware. (https://www.wondernerd.net/blog/funny-ways-to-explain-things/) In today’s post I’m going to share some on how I explain Virtualization concepts to non-techies.

Before we get started I hope this series of posts doesn’t qualify me as a Sales Person, I’d be happy with a nerd who accepts that not everyone’s mind works like mine so instead of making them think like me (which for anyone who has been around me knows thats almost impossible). So instead I’ll change my thinking to work for them. As always: They may help you they may not. You may even find them funny. If they’re helpful feel free to reuse them. If they aren’t helpful or funny please blame congress. >:)

 

Once I killed a six-pack just to watch it die:

My wife had a hard time grasping what I did when I went out and installed VMware a customers sites. So I explained Virtualization in the following manner.

Lets start with a bottle of beer. This bottle is my ESXi host, its a big computer, that’s all it is. We will use this as a container to put stuff in sort of like we use a beer bottle to put beer in. What about the beer in the bottle you may ask? Well the beer is my virtual machines or VM’s. Really liquidy and sloshes around in the bottle. The beer like our VM’s have no real shape, they are hard to hold, and can be poured from container to container really quickly. (They also happen to be what everyone wants to consume.)

During my installs I help my clients become brew masters. I show them how to bring together all the ingredients they need to make the sort of beer they want to drink. Every site has different things they want in their beer. The same is true for their virtual environment. They take different things that are physical assets and we turn them into something fluid. IE a virtual machine or a VM.

So we’ve got all these bottles of beer or physical servers with VM’s now what? It doesn’t seem like we’ve accomplished much. All we did was take several systems and combine them into one? That’s not helpful if we drop the bottle all we’ve got is a big mess!

You’re absolutely right. We need something to keep all of our bottles together and protect them from falling. Well we have this. Its the vCenter for the virtual environment its call a six-pack to most beer drinkers. Like the six-pack sleeve the vCenter holds all our bottles (errr umm ESXi hosts) together in a single package which makes it easy to carry them.  It also helps to protect them. If for some reason we break a bottle and all of our VM’s spill on the floor, our vCenter six-pack has a built-in mop called High Availability or HA.  Our mop will just soak up all the beer and pour it into a different bottle. We’re then back in business! We may be one less bottle in our 6 pack but we still have all our beer… I mean VM’s.

Well that’s great, we don’t have to worry about breaking bottles. But what if the brew master puts too much beer in a bottle or to many VM’s on a single host?

We have that taken care of with our six pack too. We have a distributed resource scheduler (DRS) that watches the bottles and keeps them from getting to full. If one bottle starts filling up with beer the tap shuts off and redirects the beer to a different bottle.

And there you have it a six-pack of virtualization goodness. Drink up!

 

and I-
I took the road less traveled by…

Explaining how multi-pathing works is hard to explain. Now try explaining it to someone who is of the old guard (someone who is stuck thinking networking is still done like it was in 1997) or is non-technical. I used the following example to explain the benefits of multi-pathing.

We’re going to start by building a standard vSwitch with 2 network connections into it.
Why two network connections?
So we get better speed.
How does that work, aren’t they just for fail over and redundancy?
Not quite.
Well then don’t we need to make them a pair of LAG ports?
Nope.
I just don’t get it!

Lets back up a little and start with a story from my child hood. (Really its not that bad!) I grew up in a small town on a gravel road. You’d be lucky if you could fit a car going each direction down the road. At the end of the road it turned into a two lane street with a speed limit of 30 miles per hour. I could take that two lane street to the edge of town and hop on a two lane state highway, it had a speed limit of 60 miles per hour. If I took the highway I’d wind up on Interstate 70 which is 4 lanes and has a speed limit of 75 miles per hour. From there I could go anywhere and get there fast!

See it was dull but it wasn’t that bad now was it?
So whats that story have to do with multi-pathing?
I just described your network. That slow gravel road in front of my house, well that’s like the 10base2 network you have running to that old box with the 5 1/4 floppy you use to run that application that they stopped making 15 years ago. And that paved road, that’s the 10/100 network that goes out to all the desktops. Its fast enough to do what the end user need to do. And of course that state highway. Its a gigabit link from those switches back to the data center.

Now let me guess your data center only has gigabit in it? So that interstate can’t be 10 gig can it?
Yeup, you loose!
You’re right! The interstate isn’t going to let me go much faster than I could on the two lane state highway. But the interstate lets me move more efficiently. 
How?
That’s simple, the highway is nice and flat, with long curves, and relatively few on ramps which means I can just drive and don’t need to speed up or slow down.
Now what if I run into the little old lady who can barely see over the steering wheel?

This is where we can utilize multi-pathing, I just move over to the other lane and go around. I can do the same thing here with your virtual environment. We’ll take this iSCSI network and set it up so that when one of your VM’s requests something from storage it can see which lane is the slowest and hop in the other one that has fewer cars in it. This is actually the idea of least queue depth.

Well that makes sense but what about all that jumbo frame stuff you setup? That doesn’t seem like it would help all that much?
Well it actually does. You’ve seen the big rigs rolling down the interstate with all the cars stacked on them right?
Of course I have!
Why do you think the automakers load all those cars on the trailers?
Well its easier to move them from one place to the other. You use less fuel and put less wear and tear on the cars. Plus it only takes one person to drive all of them across the country instead of one for each car.
So its more efficient is what you’re saying?
Yes, of course its more efficient!
You just described jumbo frames. They make it more efficient by moving more data from one place to another.

This explanation generally works, and customers seem to be happy when they start explaining how it works.

As always more to come.

 

Permanent link to this article: https://www.wondernerd.net/blog/funny-ways-to-explain-things-part2/