-1
$\begingroup$

My sci-fi story has this girl, let's call her "Lucy". She finds herself virtualized into a simulation within one or more supercomputers, which translate programs and inputs into visual cues; this processor is a room, this program is a robot, this data is a stream and so on... At first she explores her new environment, while her real body lies in stasis where she left it. The problem is the computers are under attack in the real world, forcing Lucy to transfer her data from one computer chip to another or she'll end up dead (or in this case deleted). She "jumps" metaphorically from a computer to another until she gets to safety, creating tension is the scene. That's pretty much the gist of it.

However there's a flaw in the premise: Can't she just transfer to the last computer right away?

The computers are connected so this is possible and made easy since she has master access to the computers systems (and an instruction manual, because those are important!) she could just deus ex machina her way out of the problem. The system needs to be believable or at least have a reason to be the way it is.

I'm looking for a reason one would have to link computers in a sequence instead of all together. Arguably the latter would leave the floor a mess of mangled cables to be every electricians nightmare. Or the computers being advanced and futuristic (with neon glow!) can just transfer data wirelessly. Either way what would be a good reason to have them designed this way?

Lastly while I appreciate your expertise I did not come her to get lectured about something I know little about. I have had some experience with engineering and coding before, but only ever once.

Improve this question
$\endgroup$

4 Answers 4

Reset to default
5
$\begingroup$

It is a mesh network, supporting software defined networking to implement arbitrary network topologies. Nodes can be added or removed at will, and there's no notion of a "trunk" or "master" node upon which everything (or many things) depend. This makes the system extremely flexible in the sorts of workloads it can run, and the sorts of places it can be deployed, and quite robust to the failure of many nodes whilst still being able to route around the damage.

This sounds like a way in which you might simulate (or implement) a hive mind, if each node were capable of running the regular kind... maybe this isn't the job it is being put to right now, for various plot-related reasons, but to offset its cost it is used for playing games instead.

Improve this answer
$\endgroup$
3
  • $\begingroup$ This is a better way of stating my answer. $\endgroup$
    – jdunlop
    Oct 22, 2021 at 19:30
  • $\begingroup$ (Really, I should wait for Starfish to post an answer. Saves keyboard wear.) $\endgroup$
    – jdunlop
    Oct 22, 2021 at 19:40
  • $\begingroup$ @jdunlop its very kind of you to say so ;-) $\endgroup$
    – Starfish Prime
    Oct 22, 2021 at 21:02
1
$\begingroup$

"Can't she just transfer to the last computer right away?"

The "last stand" supercomputer does not need to be defined at the start of the story

Even in a totally connected network of computers, if Lucy can habit only one computer at a time, her "relocation path" is linear.

Yes, she could relocate herself in any of the supercomputers. If they are the same or similar, there's no need of an a-priori definition for the "final computer", her "journey" would not lose any drama in the absence of a pre-defined "supercomputer of the last stand".

Improve this answer
$\endgroup$
1
$\begingroup$

She doesn't know which computer is her "final computer"

If there is a categorical "final computer" already determined by the setup, then she needs to identify it. I'm assuming that this final computer is on a local network and not escaping over the internet, which would make getting there much easier

Computers on a network can be identified by hostname (assigned by an admin) or by IP address (assigned manually by an admin or automatically by software. Either way, it doesn't identify which one is safe. Consider an admin setting up 30 machines at once. They're going to use a script to do so and so the hostnames are going to be something like "lia_seattle_01" and "lia_seattle_02" and so on.

Likewise, the IP address will either have been assigned by the admin, and so probably counts up with the hostname suffix, or is automatically assigned, and therefore random. Therefore, there's no distinguishing characteristic that would lead her to the final computer.

This means she has to load herself onto each machine to see if it has the characteristic needed for it to be the final computer.

Improve this answer
$\endgroup$
0
$\begingroup$

Networking

If you run something akin to

tracert google.com

on your computer, you'll see a series of intermediary nodes between you and Google's nearest, most convenient server. Why don't you just connect directly to Google's server? Because there isn't a wire connecting your computer to Google's datacentre.

It doesn't make sense (and doesn't scale) that every computer, even in your hypothetical scenario, be connected physically to every other computer. If "Lucy" can't see the shape of the network beyond immediate connections, she would have to send her entire self to the "next" computer and then reinstantiate herself and look around.

Now, the risk would be that an interstitial computer on the network would just be there for routing and wouldn't be able to instantiate a consciousness, but you can get around that with other worldbuilding. But not having every computer connected to every other computer is just standard operating procedure for computer networking, not something that needs an extraordinary explanation.

Improve this answer
$\endgroup$
10
  • $\begingroup$ All the hosts you see in a trace route are routers, except the last one. All of them. And they do not accept connections. They just don't. They have no reason to. They exist only for the purpose of routing packets, and are totally dedicated to that purpose. $\endgroup$
    – AlexP
    Oct 22, 2021 at 19:19
  • 1
    $\begingroup$ @AlexP I understand that. But they do not have to be. We're talking about an old-style computer fiction where individual computers are running single programs, so we get to rewrite some of how one would set up a network, and it could as readily be an ad-hoc mesh network limited to adjacent nodes, rather than how the Internet works. $\endgroup$
    – jdunlop
    Oct 22, 2021 at 19:19
  • $\begingroup$ @AlexP - I was using tracert to illustrate the fact that there are intermediary computers, and that a "direct connection" to another server is not direct, but routed through other computers, whether or not they accept connections on most protocols. $\endgroup$
    – jdunlop
    Oct 22, 2021 at 19:23
  • 1
    $\begingroup$ @AlexP - also, a minor point, but the first hop on my traceroute is my home anti-ad solution, which is a Linux box handling local DNS, DHCP, and running DPI to drop/replace ad content... and it will absolutely accept a connection. So, no. Not all of the hosts in the traceroute are only routers. $\endgroup$
    – jdunlop
    Oct 22, 2021 at 19:38
  • 3
    $\begingroup$ @LiveInAmbeR - no offense was meant, but the objection posed to my answer was "that's not how networks work IRL", and your computer world as laid out is very much "that's not how computers work IRL". Which is totally reasonable, it's a well established trope, but as you say, you're not particularly familiar with computers, and it's not integral to how the story is meant to go. The concept of networking would be sufficient, rather than drilling down on how the internet backbone routes packets. $\endgroup$
    – jdunlop
    Oct 22, 2021 at 20:06

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .