Welcome! These forums will be deactivated by the end of this year. The conversation continues in a new morph over on Discord! Please join us there for a more active conversation and the occasional opportunity to ask developers questions directly! Go to the PS+ Discord Server.

Hacking and downloading Spacehip AI's?

4 posts / 0 new
Last post
Neutrino Neutrino's picture
Hacking and downloading Spacehip AI's?
I wonder if there is some good ideas on hacking ship AI's, which is briefly covered only in the rules? Questions I am considering to tackle: - The PC's may want to 'hijack' or 'save' a particularily good or reliable ship AI. Which way is best, maybe inly possible with the permission of the ship AI? - Level of security for different ship AI's? - Ship Mesh vs. Planetary Mesh? How do you separate when you enter orbit, or are located inside a ship hangar? - Ship Mesh when in outer space, what functions and resources are available, can the Ship AI be turned on/off from the Ship Mesh as to aviod Hacking by curiuos PCäs or others? Thankful for advice,
nezumi.hebereke nezumi.hebereke's picture
Re: Hacking and downloading Spacehip AI's?
I suspect most meshes will be broken into several layers - Public layer, where anything can connects. This is like the public internet. Private layer, which has protected data and limited access. This is like private web sites and networks, protected by firewalls and other basic controls (and can normally be hacked remotely). Secured network, which is like a private layer, but has much better controls, including some which may not be crackable by normal means (for instance, QE networks). Segregated networks, which are disconnected from the mesh, or have only limited connectivity. These are the most difficult for most crackers to attack. The more critical the system, the lower on that scale it will be. I imagine a ship will have all of these layers, but the ship's guidance AI will be on a segregated network supported by dedicated connections, to reduce the risk of viral infection and so on. A possible network layout would have the sensors feed through a DMZ (Demilitarized zone, a firewalled server with very high security and limited connectivity) to the guidance network. The sensors would also have a one-way feed to the general ship systems, so the general public can get data from the sensors, but they can't send data to the sensors and up to the guidance systems. There are two communications systems, one on the public level, and one connected via DMZ to the guidance AI. Send and receive are possibly delinked, to reduce vulnerability to attack, and they only accept connections to known agents (so a random hacker can't connect to the security comm links and attack that way). At least, that's how I'd lay out a "good" ship. An average ship is probably much simpler (and much more vulnerable to attack, with only a few select systems running on segregated or hardware-only systems, such as the generator and engines). Once you have access to the AI, you have to overcome its security. Good ship AI is going to be worth a lot, and copying it creates security vulnerabilities, so it will resist copying. This may use things like encryption, copy protection, critical pieces encoded solely in hardware, or other ingenius schemes. If you have access to the security network and the appropriate skills, you should be able to rip a copy, given time. Otherwise it would take a pretty fantastic test to get the bulk of the AI (if it's possible at all), and you'll have to repair it later, if it's possible. The mesh is called the mesh because it automatically interlinks (unless intentionally segregated). So your ship's public mesh and the planetary public mesh will interconnect if within range. Once you leave range, they separate. Segregated meshes will never connect, so no difference. Parts of the ship's AI may be turned off, for instance communications. Most functions that are vulnerable to hacking will rarely be turned on (like the control panel or maintenance access). But for the most part, if a part of the AI is critical enough to be part of the guidance AI and on the segregated network, it's probably there for a reason, and its state won't change much between orbit and deep space.
Neutrino Neutrino's picture
Re: Hacking and downloading Spacehip AI's?
I like the idea and it makes sense of a 'guidance network' for ship AI's and also of use for astropilots, astronomical observatories, military, customs, travel agencies, astronomers, etc. Like a gigantic Google Earth system for space... Wonder how this 'guidance network' is maintaned and supported, since I guess it must be connected with numerous recievers/transmitters in space, and I guess it is accessed via the Mesh? Could this mean that you can search and screen for ship AI's using this system in order to find out where they are going, coming from and find out details about their cargo and passengers...? Another thing, if I don't have time to hack and copy the hard drive of the AI, I could possibly bring the hard-drive itself physically to hack it at a later stage? This may also be useful if I suspect that the AI is infected with an Exsurgent virus...?
nezumi.hebereke nezumi.hebereke's picture
Re: Hacking and downloading Spacehip AI's?
Neutrino wrote:
Wonder how this 'guidance network' is maintaned and supported, since I guess it must be connected with numerous recievers/transmitters in space, and I guess it is accessed via the Mesh?
I was referring to a specific guidance network on the ship only. It would be segregated as much as possible to avoid it being manipulated into crashing the ship into something, wasting fuel, or toasting a habitat with its engines. However, there is definitely going to be a navigation network across habitats and ships. This would be like a fan-maintained google maps, indicating which habitats are where and what they have to offer. Most ships would want to keep themselves updated (or might not be able to prevent it! With a thousand passengers tweeting every five minutes, that'll be a pretty loud mesh blip). Using network traffic, heat and engine burn signatures, officially posted flight plans and so on it would be trivial to track most ships and their courses. You can calculate the type of engine, mass of the ship, and most of the passengers, and this information cannot be easily hidden (or in some cases, hidden at all). Cargo would probably be officially declared in most cases, but not necessarily. I don't doubt that most cargo, being wirelessly active, can be 'hacked' for inventorying without too much trouble. But that would be the easiest thing to hide when shipping stuff.
Quote:
Another thing, if I don't have time to hack and copy the hard drive of the AI, I could possibly bring the hard-drive itself physically to hack it at a later stage? This may also be useful if I suspect that the AI is infected with an Exsurgent virus...?
With segregated networks? Maybe. A lot of data will be concentrated on a few devices, but in the mesh-age, you can expect data and functionality to be distributed across multiple nodes. So perhaps trajectory calculations are stored with sensors, fuel calculations in the fuel storage area, maps in the guidance area, and it all comes together into a single user interface, in which case creating the whole AI would involve pulling pieces from these different areas and reassembling. This also precludes active security controls, such as pieces of the AI being hardwired into separate modules (either to prevent tampering and viruses, or as copy-protection). As a GM, the details are really up to you.