Blog

hero-resources.png

Hybrid Cloud Economics: Does the Math Work Out?

Allon Cohen August 22, 2017 blog, hybrid cloud, cloud economics

By the end of this blog I will:

  1. Use macroeconomic principles to prove that every growing enterprise – yes, everyone – should use hybrid cloud services, and
  2. Show a nice hybrid cloud demo.

If you just want to view the demo click here. If you want the proof that everyone should use hybrid cloud, read on!

The Economics of Hybrid Cloud

While the economics of public cloud usage are relatively new, the underlying economic principles are surprisingly similar to 19th century macroeconomic theory. In a previous blog we stumbled upon the politically sensitive subjects of public cloud and land ownership. In this blog, we blatantly repeat our political insensitivity, by discussing the similarities between hybrid cloud solutions and the principles of free trade.

Macroeconomic specialization is a much-detested subject, mainly because, as economics students will testify, it is used by unscrupulous professors to create easy lures to wrong answers in multiple choice questions. However, it is also hated, by some, because it is the base of the economic proof that free trade brings value to everyone. Yes, everyone, even those that have an absolute advantage in producing everything.

I’ll get to cloud shortly, but to explain the principle we need to look at a simpler example first.

Consider the example of Batman and Alfred. 

It’s evening time, Batman and Alfred have 10 minutes and need to decide who should mix cocktails and who should make the sandwiches.

 

Make Sandwiches

Mix Cocktails

Batman

Makes 6 sandwiches in 10 min

Mixes 2 cocktails in 10 min

Alfred

Makes 2 sandwiches in 10 min

Mixes 4 cocktails in 10 min

Who should do what? Yes, it’s an easy question, but we are just getting warmed up. Batman has the absolute advantage in making sandwiches but Alfred (a trained bartender) has the absolute advantage in mixing cocktails. If Batman focuses on sandwiches, and Alfred on mixing the cocktails, and they trade in the goods, everybody wins. This is a simple result, and was widely known since at least the 18th century.

Now let’s turn it up a notch. Morning time comes, and Batman and Alfred need to decide who should scramble eggs and who should brew coffee. Yes, a superhero’s day starts with breakfast and they must make a decision.

 

Scramble Eggs

Brew Coffee

Batman

Scrambles 9 eggs in 10 min

Brews 3 cups in 10 min

Alfred

Scrambles 4 eggs in 10 min

Brews 2 cups in 10 min

Batman has been working out, and is faster at both scrambling eggs and brewing coffee. Having the absolute advantage, should he do everything? Should they split the time and each do a bit of both?

Here is where 19th century economist David Ricardo’s theory kicks in. It looks at comparative advantages.  You see, in the time that Batman would spend making 1 coffee cup, the duo would lose the ability to potentially scramble 3 eggs (because this is the amount of eggs Batman could have scrambled if he didn’t waste time on brewing one coffee cup). Remember, he can make 9 scrambled eggs in the same time he makes 3 cups of coffee, a 3 to 1 ratio.

On the other hand, if Alfred brews a cup of coffee, the team only loses 2 eggs! (Alfred can scramble 4 eggs in the time he brews 2 coffee cups – a 2 to 1 ratio).

So, the team loses just 2 potential eggs when Alfred makes coffee instead of losing 3 eggs when Batman makes coffee (in economic terms the opportunity cost of Alfred brewing coffee is lower).

Aha! Alfred should be the one focused on making coffee, because overall the duo gains more eggs that way! Batman should steer away from making coffee, not because he is bad at it, but because the team can use his talents in a better way…scrambling eggs.**

This principle of specialization, whereby it is comparative and not absolute advantage that matters, is one of the most counter-intuitive results of classical economics. But the math works out, and it has spurred global growth and efficiency of production for some 200 years. Each person/factory/country/data center focuses on the area it has a comparative advantage, and everyone wins.

There is an important caveat, however. The end results must be freely traded. It only works out when Alfred and Batman can share in the results of their labor. Batman can provide Alfred with some of his scrambled eggs, and Alfred would be kind enough to pass Batman a cup of coffee.

The Economics of Hybrid Cloud

Armed with this insight we go back to discussing the economics of hybrid cloud deployments.

You see the real Batman of the enterprise world is your internal, on-premises, IT department (yes, I know, flattery gets you everywhere). The real Alfred of this world (much like his friend Siri) lives in the cloud.

And yes, like Batman, internal IT has an absolute advantage over public cloud in (almost) everything. But the cloud has a macroeconomic trick up its sleeve. The cloud has a comparative advantage.

Consider a company that needs to do two basic IT tasks. Run an enterprise application, and run analytics on a big data application.

Let’s assume that running both applications produces value to the company (otherwise why run them?).  The value generated per server, looks something like the example below:

 

Value generated from running Enterprise App

Value generated from running Big Data App

On-premises IT department (Batman)

$900/server/month

$300/server/month

Public Cloud (Alfred)

$400/server/month

$200/server/month

Yes, this is just an example, and the numbers might look familiar (keeping the example simple). But there are few assumptions which I am using:

  1. I’m assuming running the enterprise application on-premises generates more revenue than running the big data applications. The enterprise application is at the heart and core of the enterprise value creation. The big data application is analyzing data generated by the enterprise application.
  2. I’m assuming running both applications on-premises would generate more value per month, because right now for many companies, cloud is more expensive (on absolute terms). This may change in the future, but let’s keep it that way for now.
  3. I am assuming the ratio between the two is lower in the cloud than on-premises. This is the comparative advantage. There are applications, like analytics, that even if on absolute terms it is more expensive to run them in the cloud, on comparative terms, the cloud runs them more efficiently. (But also in general, between any two applications, there will always be a difference in the ratio that will force specialization in one direction or the other!)

The result?

I’ll let you do the math, but unsurprisingly, the principles of macroeconomic specialization hold. Optimally each application should be run in the location which has the highest comparative advantage. When a comparative advantage exists, you should use the on-premises servers to run your enterprise app, and you should use the cloud servers to run the analytics application.

Why not run the analytics application on-premises? Because it is a waste of these precious on-premises servers! Use them to run the enterprise application instead, and you gain more value. Each time you take an on-premises server and switch it to running analytics you are losing value, because of the opportunity cost. Hybrid cloud is your way to gain the most value.

Why not buy more on-premises servers and run the analytics on them? The answer is simple: Yes, you can buy more on-premises servers, but it’s better to use the new on-premises servers to continue to run more of the application that has a comparative advantage on them. As long as you are growing, you would still gain more value with a hybrid approach using the cloud for its comparative advantages.

Is there a Caveat?

Yes, like in the Batman example, for all of this magical value creation to happen, Batman must learn to share the results of the efforts. No, you don’t need to pass coffee and scrambled eggs between on-premises and cloud environments, but you do need to share the data.

In the example above, the data generated by the on-premises app must be shared across the cloud boundaries for access by the analytics application. Needless to say, the data generated in the cloud must also be usable by your on-premises team.

Seamless data movement across cloud boundaries, is the equivalent of free trade for the hybrid cloud world. It is the engine that drives the macroeconomics creating value for everyone, yes everyone.

Oh…and here’s a demo of how it is done with Elastifile’s cross cloud data fabric.

 

**Technically, Batman should only start making coffee if the team is sick of all the eggs he has scrambled and needs more coffee than Alfred can produce.  But we’ll save that discussion for another blog.