My Last Expensive ComputerWhy 2013 was the last year I'd buy an expensive computer.

How I plan to save at least $400 a year by no longer buying high-end computers to handle my intense computing workload.

Moving To The Cloud

February 2013, I bought the latest version of the MacBook Pro (Retina) at the time, fully decked out with the exception of the storage space for just over $3,200. Fast forward 18 months later; you can now buy that very same laptop brand new on eBay for $2,300.

Let me do the math for you; that’s a depreciation value of $50/mo.

For $40/mo on DigitalOcean I can get a virtual server with 1/4 of the memory and 1/2 of the cores of my MacBook Pro. You may be thinking: “It’s lower-end hardware, so it’s going to be cheaper captain obvious. How does that make any sense?”. As you read on, I’m going to explain how you can get a more powerful droplet than the MacBook Pro I have and still wind up paying less.

Questions that I had to ask myself:

  • Are you really using your computer to its full capacity around the clock?
  • How often are you using your computer to its full capacity?
    Maybe a month out of a year.
  • How are you using your computer to its full capacity?
    99% of the time it’s due to heavy data processing.
  • Aside from design/engineering, what do you really need your computer for?
    Surfing the web, streaming media (music/videos) and reading books. (Really? I couldn’t believe it either.)
  • If you had access to a supercomputer, how much time would you have saved by not waiting on those operations to have completed on your computer?
    I didn’t have a definitive answer, but I imagined it would be significant.
  • If you moved everything to the cloud how much would you save a year?
    I’ll answer this question below in detail.

All of the questions above were extremely critical for me coming to a final decision as to whether or not I would make the full move to the cloud. Having a hardcore beast of a machine sitting in front of me is something that I’ve grown accustomed to; removing that, well let’s just say I’ve had easier decisions.

How Much Would You Save?

A few things that you need to understand before you proceed:

To effectively pull this off, you’ll need to destroy and save a snapshot of your droplet when you’re not using it. Then, when you do need to use it, you’ll have to create a new droplet from the snapshot that you previously saved.

To get a realistic estimate of how much you’ll use your computer in a given year, you’ll really need to track your usage. How many hours are you on the computer during weekdays, weekends, holidays, vacations, etc?

I also realized that there may be times where I’d need a high-performance droplet to do some hardcore computing outside of the norm, but the majority of the time the bare minimum would do just fine.

Applying the cheapest option available combined with infrequently needing a higher performance droplet, would cause the price to balance out making the $40/mo droplet a realistic baseline cost.

Here are my results for the amount of time that I’d be cloud computing in a given year:

  • $40 a month for the droplet.
  • 12 hours a day Monday through Thursday.
  • 8 hours a day on Friday.
  • 4 hours a day on Saturday and Sunday.
  • If I spend 3 weeks on vacation each year, that’s 21 days that I wouldn’t be performing any cloud computing.

Formula: ((12 * 4) + 8 + 4 + 4) * (52 – 3) = 3,136 Hours of Usage

That’s 3,136 hours a year costing me a total of $200/yr (overly-estimated) for both my droplet and snapshot image!

  • CLOUD COMPUTING: $200/yr
  • TOTAL SAVINGS: $400/yr

Again, this is just my personal figure. Most engineers I know (especially those with families/children) wind up spending a lot less time than on a computer; making the savings even higher for them.

Cloud Cost Calculator

The following form will allow you to calculate just how much it would cost you to move to the cloud on a yearly basis.

Click here to see DigitalOcean droplet prices.

Looking Back

Currently, there’s no easy method of destroying and rebuilding droplets on the fly without having some sort of engineering background. Surprisingly, I think you’ll find that even if you were to leave your droplet running around the clock, you’ll still wind up with savings come the end of the year.

If you’re interested in learning how to programmatically destroy and rebuild droplets, check back on the blog as I’ll be posting some¬†DevOps¬†tutorials illustrating how to do so.

About Salvatore Garbesi