There and Back Again - a Photo's Tale
Sarah and Miguel, along with their band of friends, stood at the Sun Gate, catching their first glimpse of Machu Picchu as morning mist lifted from the ancient stones. Four days of trekking the Inca Trail had led to this moment. Sarah held up her phone and gathered everyone together, capturing the whole crew grinning like idiots with the ruins stretching out behind them. Perfect. She slipped the phone into her jacket pocket.
That afternoon, on the train back to Cusco, exhausted and half-asleep, she woke to Miguel shaking her shoulder. Her jacket was gone. The woman across from them? Also gone.
She spent the rest of her Peru trip phoneless. Fortunately she and Miguel still had the information and tools they needed to complete the few days left in their trip before they flew home.
Once back in Portland, Sarah went to her local carrier store and got a replacement phone. When she returned home she configured her new phone to replace the old one and went out for a walk. By the time she returned, there it was, along with her contacts, calendar, email, and so much more. There was that photo—the whole crew at Machu Picchu. There. On her phone. As if it had been there all along. She immediately posted it to The Gram; better late than never.
Like Magic. Or something indistinguishable from it.
Fellowship of the Heights
To consumers, the cloud is invisible. You don't think about where your photos live, they just exist. Gone are the days of navigating to C:\Documents and Settings\Sarah\My Pictures or /home/users/sarah/photos. Now there’s no filename, no directory, no extension. Just a thumbnail. Tap it, see the photo. Take a picture on your phone and it appears on your tablet, laptop, desktop, smart TV, you name it.
But let’s rewind to that moment at the Sun Gate and discuss what really took place.
When Sarah took that photo at Machu Picchu, her phone had no signal to upload it. However, that afternoon on the train back to Cusco, somewhere during those two hours while she slept, the train passed through areas with enough cellular signal to connect to the internet. Her phone, working silently in her jacket pocket, was able to get our hero photo uploaded before the bad actor made off with her possessions.
That image traveled from a moving train in the Peruvian Andes, over the airways, routing into Latin America, and then possibly under the seas to the United States, until it reached a collection of buildings packed from floor to ceiling and wall to wall with computers. These buildings are called data centers, and they are where “the cloud” actually lives. Sprawling real estate packed with equipment and cooling and very VERY off-limits to the average person.
In all likelihood the photo wasn’t just stored once. Depending on the application or service, that photo was probably written multiple times to distinct locations; either separate buildings in the same location, or possibly to buildings in different locations, separated with the intent to be resilient to natural disasters, electrical disruptions, weather, and other factors.
When that thief lifted her jacket on the train, they got her phone. But Sarah’s data was already backed up securely thousands of miles away. Back home in Portland, when she obtained a replacement phone and logged in to her app, the software on her phone talked to a service in the cloud, just as it did when the photo was taken, and hardware behind the application knew what to retrieve and where to retrieve it.
Starting to sound a little less magical?
The Two Clouds
Perhaps you've heard of fractals before. The more you zoom in on a fractal image, the more detail it reveals. The same could be said for a cloud. From a distance it's a fluffy shape, but up close it has depth and structure you couldn't see from afar. The clouds that span businesses and the internet work the same way, the more you see the details the more details there are to see.
So here's what keeps infrastructure engineers up at night: all those layers of the cloud. All the services, all the connections between and within them, all the physical machines that power them, all the data within them... there's immense complexity. Complexity within, but from the outside, something mundane and, while appreciated on occasion, mostly ignored.
In this example, an application on Sarah's phone connects to a service in the cloud. That service knows who Sarah is when she logs in and keeps track of things like passwords, billing information, and other "account" things. More important to our story, however, that service keeps track of all the photos Sarah has taken. It has a list of filenames, often seemingly random numbers and letters, that belong to Sarah. That list is Sarah's saving grace.
That photo may have a bizarre and nonsensical name to a person (if people ever actually see the names of files anymore) and as such it doesn't exist anonymously in some abstract cloud. It exists as ones and zeroes stored on specific hardware in specific locations. Rather than a file on a hard drive, it's a file in the cloud, and is stored as an object in a service that may be unique to Sarah's photo app with millions of assets, or may serve many applications and have billions.
These object storage systems are massive collections of simple computers with huge arrays of hard drives. They are all managed by an algorithmic model of storage that allows the objects (files) it stores to be distributed neatly across many different machines both by quantity and size, reducing the "hot spots" where failures may occur, lengthening the useful life of the equipment, and ensuring resilience to problems like drive failures, power outages, etc.
When the photo service needs Sarah’s photo to rebuild the collection in her application, it asks the file service for "object xyz". The file service figures out where the file lives (e.g. which physical machine has the data) and retrieves it. Two photos Sarah took one after another might actually be stored "in the cloud" on different machines in different buildings, but she doesn't know that, her app doesn't know that, the photo service doesn't know that, only the file service does. That's part of the beauty of the cloud; the localization of what matters and where allows for blazing speed.
Return of the Bytes
Think about it. When Sarah got her new phone, turned it on, and logged in, the photo app:
- Asked the photos service to find all photos belonging to Sarah
- The app then requested each of those photos from the service
- The photo service asked a file service for each specific object
- Each photo had to be located on a machine in a building
- The file was returned by the file service
- The photo was delivered by the photo service
- Sarah saw the photo appear on her phone
One corrupted index entry, one failed hash lookup, one unreachable storage cluster, and the magic breaks. The photo doesn't appear. The illusion shatters. This is where redundancy, checksums, backups, and other resilience techniques minimize as much as possible (there is no 100% guarantee) the inconvenience to the end user, despite the mountain of process and information required to do something as simple as store a photo online.
Arthur C. Clarke once wrote that any sufficiently advanced technology is indistinguishable from magic. To Sarah, Miguel, and their friends, that's exactly what happened. Their photo survived because of magic—the benevolent magic of "the cloud."
But the truth is more remarkable than magic. It’s technology they’re blissfully unaware of, working in perfect coordination to save not just that moment on their phone, but replicate it across the world before they ever had a chance to share it.
The real miracle of the cloud isn’t that storage is infinite or ubiquitous. It’s that an inconceivable number of specific, precise operations happen in perfect coordination, at scale, reliably enough that we call it “magic.”
To the consumer: invisible, automatic, simple.
To the business: a carefully orchestrated symphony of specificity, where every byte has an address, every file has a location, and every action requires coordination across thousands of physical systems.
The photo of Sarah, Miguel, and their friends found its way home not because the cloud is magical, but because thousands of engineers built systems specific enough to find one file, on one drive, in one data center, and deliver it back across continents—in a sea of billions.
That's the real magic.
While Sarah's photo recovery may seem magical, behind it are decisions about storage, replication, resilience, and cost. I've been helping make those decisions for 30 years—building applications from startups to enterprises, operating at scale across data centers, colo facilities, and cloud platforms.
If you're interested in optimizing cloud spend, migrating workloads, or modernizing systems, my team and I can help. Learn more at thesteveco.us.