CFD3 Prep Post: Morpheus Data

This is part of my series of posts on Cloud Field Day 3.

Morpheus Data bills itself as a “unified multi-cloud orchestration tool”. Okay. What is that, exactly, and why would I want one?

It’s a kind of manager-of-manager technologies that sits over the top of all your other automation tools and coordinates their actions, essentially. It connects to Chef, Puppet, Ansible, Terraform, Docker, SaltStack, GitHub, Jenkins, etc., etc. and orchestrates all these tools to they work together to achieve an outcome. That outcome might be “Create a new database instance in AWS and deploy the latest code build”. This action might get triggered by a new build being tagged in your source repo, or maybe it’s done by someone clicking an icon in a GUI somewhere.

The theory is that it’s easier to use a tool like Morpheus to do this than to write your own integration scripts. Morpheus apparently talks to a bunch of platforms and tools already.

The headline on the Morpheus website is that it’ll save me money and time, which is fine, but everyone says that, about every single technology. Everyone claims some sort of saving anywhere from 10% to 90% and some astronomical speed gain.

Technology Gets Cheaper

Here’s the thing about technology: it gets faster and cheaper over time. You might have heard of this thing called Moore’s Law once or twice. A bunch of other stuff has also gotten faster, denser, or cheaper over time:

• Here’s the cost-per-gigabyte of storage over 35 years to 2014.
• Here’s a graph of the price of flash memory over time.
• Here’s the price of light (as in, from lightbulbs) over time.
• Here’s prices for a bunch of different technologies over time from the US Bureau of Labor Statistics. Cable, TV, and radio service is the only thing that went up.

I highlight this to show why a price saving for a new(-ish) technology is nothing special.

What can be special is a saving in labour, which is where the “saves you loads of time” claims come in.

Labour Gets More Expensive

Labour costs tend to rise over time, outside of shocks like recessions or depressions. Here’s some data on US Labour costs since 1950:

Source: tradingeconomics.com

It goes up, but relative to the drop in cost of a megabyte of storage, not very much. That’s part of why we have computers in everything now, and why billions of people worldwide have a network-connected camera with GPS in their pocket that can incidentally also make phonecalls, but avocado-on-toast costs $15.

Savings in time can be important, but the key here is in understanding why a new system will save time.

Labour Saving Devices

Some technologies are wonderful in that they replace human labour in achieving a task, so we can go do other things. Washing machines, for example, mean that we don’t have to spend hours manually scrubbing clothes on a washboard and wringing them out (or using a mangle). We can use that time to work on our startup that turns electricity into Ponzi-schemes, or watch people play computer games on the Internet instead.

The challenge I set myself when evaluating one of these labour-saving approaches is to consider relative to what? Usually the comparison point is set to whatever you’re doing now which is an option, but it’s not a great one—in my opinion—because it ignores all of the other alternatives. That’s not how good investments are made.

Instead, I should be looking at the savings compared to the next best alternative. The concept of opportunity cost comes into play, which is the cost I incur from choosing the Not Best option. Many vendors will make a comparison of their shiny new tech with the old-and-busted tech you currently have, which isn’t fair. Of course the new shiny is faster, because as we saw above, that’s what happens with technology. Any new computer (or SAN, or router) bought today will be faster than one you bought five years ago.

Being better than what you already have is the lowest of low bars for changing anything, because if it made things worse, changing things would be stupid. You should be checking to see if there isn’t something else out there that—all other things being equal—isn’t better than the new option you’re considering.

When looking at a solution that claims to save me time, I should similarly be comparing it to alternate methods of achieving my goal, not what I’m currently doing.

Stop Hitting Yourself

“There is nothing so useless as doing efficiently that which should not be done at all.”
— Peter Drucker

The single greatest option that people overlook is the “Stop Hitting Yourself” option. Instead of spending a bunch of money on a new tool, you could simply stop doing stupid things.

For a lot of organisations, automating or orchestrating their existing byzantine bureaucracy is just layering more and more layers on top of the stupid stuff their already doing. And because it’s the same systems that built the old stupid stuff that are building the new thing, it will also end up full of stupid stuff.

Which we can then fix with another layer of orchestration! Hooray!

What I want to see from Morpheus is if it can identify a bunch of stuff I should simply stop doing. That, all by itself, would provide loads of time and cost savings, and is honestly what most organisations need more than anything else.