1
$\begingroup$

In a galaxy where humans started colonizing it, assuming internet and databases are still the way which these humans are using to save/monitor data of the systems they need to sustain their civilization (financial, medical and etc).

How would the technology keep in sync between different planets?

For example, if there is a Galactic Map that reads from a galactic map database (inside this DB are planetary systems, stars, gas giants and such) how would it be in sync with a planet at planetary system A if a planet at planetary system B made a change in the Database (I assume that every planet have read/write privileges to this particular database and changes are made according to a procedure that allows each planet to add new data to the DB)?

$\endgroup$
  • 4
    $\begingroup$ Using database transactions, same as today? I am not sure if you are aware of the limitations of the speed of light for data transfer, or do you have a solution for that in mind and the question is really about an engineering problem? $\endgroup$ – dot_Sp0T Dec 26 '17 at 13:18
  • 1
    $\begingroup$ Currently, there's no known way to travel faster than the speed of light. That includes data transfer (of any kind). Hence, unless such way is discovered, there will be no shared (and updated) internet. On the other hand, if this discovery is made, the implementation of the system will depend on the technical aspects of this faster-than-light travel. $\endgroup$ – Laetus Dec 26 '17 at 14:22
  • 3
    $\begingroup$ I was going to suggest quantum entanglement communication, but then I read this: en.wikipedia.org/wiki/Faster-than-light_communication, which says that even that doesn't work. Nothing says your universe can't be made to let it work, though. $\endgroup$ – computercarguy Dec 26 '17 at 15:26
  • 2
    $\begingroup$ We do have a handful of similar questions already on the site. You might be interested in for example How would interplanetary stock-exchanges work? Also, note that edit conflict resolution is, to some extent, a solved problem; we have a pretty good idea of how to coalesce differing datasets while maintaining consistency, especially if we can restrict ourselves to data on particular forms rather than needing a solution that works for any arbitrary data. I suspect the big key would be to not think of the database as monolithic. $\endgroup$ – a CVn Dec 26 '17 at 15:41
  • 2
    $\begingroup$ Other questions that may be of interest might be Interstellar internet use cases, or “What if the internet was mail-order?” and End-user experience and prominent use cases of robust interplanetary internet and Information Exchange In Space, likely among many others. $\endgroup$ – a CVn Dec 26 '17 at 15:43
3
$\begingroup$

You're discussing a problem that already exists today, just on a smaller scale. It'w worth noting that there are (for the purposes of this discussion) two kinds of data.

  1. Data that comes from all over and has the potential to simultaneously affect the same data element. A ruthless example is a stock exchange. A more common example would be a retail shop.

  2. Data that really can only come from one source. An example of this is an individual's medical data. A person can only be in one place at one time, so there's no risk to their medical data running into concurrency problems when it's absorbed into, e.g., the "central database."

Both kinds of data have problems when they become massive (and we're talking massive on a scale that boggles the mind). Gratefully, the vast majority of data is archival, which means it doesn't need to be synchronized. But we do have one practical limitation to consider first.

C

As we understand physics today, nothing moves faster than the speed of light. That means, for example, we're modulating a very impressive laser to ship data between star systems. It also means data updates between star systems take years.

Let's ignore C

C is boring. You cannot have a centralized galactic government if C is involved. By the time the reigning monarch's first decree arrived at the far rim said monarch and hundreds of generations of heirs would be long dead. So, let's ignore C and suggest you have a way to transit data in a pracitcal amount of time... like instantaneously.

Not all databases need centralization

First of all, most of your data will be gathered regionally and only the important bits (e.g. summaries, abstracts, compounded affects, etc.) would be moved elsewhere. There's no need for every database in the galaxy to have a copy of Lauren Farseer's latest novel. In this regard, you need to think like a library system. The less common data is copied out to remote users as-needed rather than stored at all locations. This may increase access time, but it makes the database more practical.

Concurency needs rules

Is it fair that your bid for Omicronian Jelly Beans came in so late that the price had gone up and down a dozen times, making it no longer valid? Even in today's marketplace, stocks and commodities are usually purchased through brokers, people "on site" who will act on your behalf. It's been a while since I cared about organizations like Ameritrade, but when I was trading, even though it looked like I was directly on the floor expressing trades, I wasn't. Ameritrade was my broker and they were acting on my behalf. And according to their rules, if my order came in too late, it was ignored.

You'll need to set up similar rules. No matter how "instantaneous" communication appears to be, it really isn't. It isn't even now on our own planet and never will be. You can always slice time just a bit thinner. Rules exist in today's exchanges to ensure that this very problem can't upset the trading system. You'll need to set up similar rules.

But, bear in mind my previous assertion. What's the point of bidding on Rigelian Wheat when you're on the far rim of the galaxy? Shipping won't get it to you before your sun has turned supernova. Even at FTL... that's a long trip. So why would anyone care other than to play the market for profit? One of your first rules will likely be there is no central stock exchange, only regional exchanges, and goods/stocks for sale on those exchanges must actually be within the region before they can be listed.

And this is likely the specific answer to your question, despite my lengthy soliloquy. People on one side of the galaxy don't need real-time data on flight availability from the other side of the galaxy. Your database management will become highly localized with rules governing how that information gets outside the local regions. If, for some reason, the Monarch wants to go on a tour of the galaxy, I can imagine a "bot" the manages the itineary "as he goes," dealing with databases in local regions and pre-informing the next region in line to anticipate his requests.

Tagging data really isn't your problem

Today we can tag data with time to the microsecond. Frankly, it would be trivial to tag it with time to the nanosecond... but we really don't have a reason to do that (yet). Your future world could easily tag timem to the femtosecond, which would alleviate most concurrency problems. Tag the data as well with an "originating location" and you're done making the data element unique. This isn't at all different than how the Internet works today. We started with IPv4 (255.255.255.255) which addressed 4.3 billion computers. Now we have IPv6 that cann address 281.4 trillion computers. You just keep adding blocks of 256 (remember that computers start at 0). And just in case you're a privacy nut, you can rest assured that the Galactic Fraternity of TOR can be found in your neighbor's homes throughout the galaxy.

Your real problem is relevancy

The amount of data you're talking about is ENORMOUS. In our world today, the top-10 data centers add up to about 25 MILLION square feet. That's a lot of hard drives! So, your real problem is figuring out how to balance local needs (million square feet data centers placed around a planet) with regional solutions (that fill small moons) to "galactic" solutions that fill entire planets.

Frankly, coordinating the data is easy compared to finding places to put it on that scale. This means you'll need a spectacular effort to realize the relevance of data ... and dump what isn't.

If, on the other hand, you're dedicated to retaining every random thought that enters a teenager's mind and every photo mom takes of the family vacation... then the option is to have Death Star sized data centers floating by the millions throughout the galaxy.

$\endgroup$
  • $\begingroup$ I like this answer. I'd like to add a comment about how data storage is done today - we've pretty much gone away with relational databases as most people know them for large datasets. To fill the gap most companies have moved to Hadoop, which allows you to store information on cheap commodity hardware rather than expensive products. I imagine a factory orbiting a boring star that just dumps out computers (as they are in the future) to orbit the star indefinitely and hold just a little bit of information each. It's much cheaper than building a moon $\endgroup$ – bendl Dec 29 '17 at 3:37
  • $\begingroup$ just sitting here, as a linux admin, and immediately started worrying about the procedures necessary to do a full backup of one of the more central data-centers.. its like "oh crap USB_DEATHSTAR was disconnected before umount was done, the ext-10 filesystem is corrupt, now we hope the e2fsck will fix it or we need to insert the last dump from USB_DEATHBACKUP, which "#>ls -lah" says is 245 YottaByte zip compressed.... plain text" $\endgroup$ – Confused Merlin Jan 2 '18 at 10:52
  • $\begingroup$ @ConfusedMerlin, I about died laughing! You definitely have more insight into the problem than I do. $\endgroup$ – JBH Jan 2 '18 at 16:11

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.