Warning: This post contains potential spoilers for Disney Presents: Rogue One: A Star Wars Story. I mean, you know they eventually blow up the Death Star right? So if you don’t want to know how storage plays into the plot of the movie, stop reading. You’ve been warned.
I read a piece by Lee Dallas that reassured me I wasn’t alone in the universe. I saw Rogue One over the weekend. There has been a lot of reaction to the film. Some call it a refreshing change of tone for the franchise, other a dreary slog with unmemorable characters and an ultimately irrelevant plot. I definitely fall more on the positive spectrum of reactions. Right after the showing, I had had a shocking realization. I’ve been a Star Wars fan most of my life. I’ve read my fair share of the expanded universe novels, played most of the video game properties, and seen the movies more than I’d be comfortable counting. But seeing Rogue One made me realize that for all the monolithic terror the Empire represents, they have garbage IT.
Imagine you were provisioning storage for a large enterprise or galactic empire. You have a huge footprint. Obviously, your organization is generating a lot of data that needs to be stored somewhere (image the log files for droids). This needs to be addressed. In Rogue One the solution is basically a data planet. Okay, maybe I can understand that this is some sort of mandate. After all, I’m sure the Imperial CTO is still pretty deferential when the Emperor suggests an idea. Personally, given that the Empire has millions of planets to play with (and just to decrease latency), I might distribute the Big Evil Data Pool across more sites.
So the idea of a data planet is kind of dumb, it’s a big old central point of failure, but I can at least understand the planning might be out of competent control. But how is this data actually accessed? Oh, you probably just use a local intranet to log in securely? No? Oh, wow you really want to control access. Wait, you actually have to physically go to the data?
Yes, in Rogue One to access the super secret Death Star plans, all Imperial data is stored in a giant cylindrical array. It’s very on-prem storage. But with all that data, there must be some robust indexing and retrieval systems? Nope! There’s seemingly no searchable metadata, with the characters having to search for the specific name of the project file to find it.
But after you have to physically be in proximity to the data, and find it in an unsearchable morass, you can then easily retrieve it, RIGHT? Haha, welcome to the hellscape of Imperial IT. Just for kicks, you have to manually operate a mechanical retrieval system. In a universe with laser swords and faster than light travel (we’ve established at least .5 past light speed), I guess the automated VHS retrieval of 1995 is beyond them.
The whole installation has a giant satellite dish on top for transmitting the data, which means they must do this with some frequency. Of course, if the whole point of all of these other inconveniences is to it hard for data to get out, the whole wireless data transfer seems to defeat it.
So here’s what the workflow looks like for the poor IT schmuck whenever a request comes through.
Imperial Officer: I need the plans for the Death Star.
Imperial IT Schmuck: Ok, do you have the code name for that.
IO: Uhhh, I’m not sure, you mean it’s not just called “Death Star”?
IITS: Hmmm, I don’t see any file by that name, and I have no other way to search for it.
IO: Okay…. [thinks for a minute] OH, its “Star Dust”, good thing I remembered! What a weird name for a project file!
IITS: I know right! Well I’ve located it’s physical location, let me just walk down to the storage array and get that for you.
[Imperial IT Schmuck physical walks down to the storage array, uses a set of pointlessly manual controls, and physically pulls the hard drive.]
IITS: Here you are sir. Would you like a copy on a disk or sent to your ship?
IO: The ship is fine.
IITS: [makes copy that for some reason required the actual drive in hand, then transmits an unencrypted copy to IO’s ship] Done. [repeats process in reverse to return storage device to the array]
IO: Thanks (note, he now has a completely unsecured infinitely copiable piece of data that he can transfer with far less latency and effort)
Imagine if the staff at that installation get any kind of volume! Let alone the latency from the time of request to getting the actual data. The entire system has a massive bottleneck. There’s only one set of controls for the array.
I know Star Wars is more science fantasy than science fiction. Generally the technology is just fantastic window dressing for the space opera at play. But seeing the absolutely insane state of Imperial data storage and retrieval, it plainly shows the Empire is built on pure evil.
- Palo Alto Networks To Acquire CloudGenix | Gestalt IT Rundown: April 1, 2020 - April 1, 2020
- Two Petaflop-Scale Approaches to COVID-19 Research - March 30, 2020
- O’Reilly Closes Event Business | Gestalt IT Rundown: March 25, 2020 - March 25, 2020
- A Tale of Two Data Visualizations: Ransomware Edition - March 23, 2020
- Virtual Events, Real Challenges - March 19, 2020
- GitHub Acquires npm | Gestalt IT Rundown: March 18, 2020 - March 18, 2020
- The Storage Lessons from Westworld - March 12, 2020
- It’s x86 Vulnerability Week | Gestalt IT Rundown: March 11, 2020 - March 11, 2020
- Ampere Altra Brings 80 Cores to Servers - March 5, 2020
- The Enterprise Response to COVID-19 | Gestalt IT Rundown: March 4, 2020 - March 4, 2020