How hard is it to track the location of an agent / AI if you only have it's Access ID, assuming it is running independently and chilling in a NERPs vending machine?

Not hard at all. A standard tracking test (using the logs, if that's what you have to work with) allows you to locate the access point to the wireless area a couple dozen meters wide.

I don't know why this does not sit well in my head.
If the agent/AI used a node (for example a commlink) and left the node's access ID, sure that would easily be tracked no matter where the node went within a wireless network. But to be able to track the agent/AI implies many bad things. Like being able to search for unknown agents in an area just because they are being broadcast in some fashion. Are normal users subscribed to a node also being broadcast from the subscribed node?

Um... if something is being broadcast, isn't it trackable by definition? I can't see every single node cluttering up the Matrix by broadcasting the access IDs of everyone using it, though. That seems like something only a secure node would do, and then only locally.

With the nodal anture of hte matrix, I'd execpt the data packets to include a header with some sort of location/indetification informations (probably including the accessID) to send said information in the right direction.

Careful of the logical circle! If an agent is trackable by it's hardcoded Access ID, then it's Access ID is being broadcast. If it's Access ID is being broadcast, then it is trackable.

My question is, is the Agent's personal hardcoded Access ID being broadcast by the node it is in by default.

Broadcast is not the word you are looking for. The AID is there, and can be read, but you will not get this information unless you actually analyze the Agent.
That being said, If it is present in the Node, and a Spider/IC, also present in the node, wants to track the Agent, then the Access ID is available for such things, and the Agent can be traced.

If it is in a NODE in China, and you are in Seattle and want to find that particular AGENT's AID, then it is a bit more difficult. Since you are not in the same node, Tracking is mostly not going to happen. You need to locate the Access ID (Browse should work, with a high threshold to represent the mass amount of data you are filtering through), once you have located the AID, then you can travel to that node and try and trace the signal from there. You NEED to have a place from which to start tracking.

It is harder to do with Hackers, since the smart Hacker changes his AID for each and every run. You need to be in the same node, essentially, to trace his signal.

Anyways...

So if I am in Seattle and the agent is in my node:
1) I use analyze on the Agent and grab its Access ID.
2) The Agent then leaves.
3) After much obfuscation the Agent is loaded into a NERPS machine in China.
4) I can then use Browse, looking for it's Access ID, to find the node it is currently in. (This is a datatrail search, NOT the same as a standard tracking test, and it may be near-impossible if measures are taken to not create a trail.)
5) If a data trail is found pointing to it's current location, I then begin a trace on the NERPS machine Access ID to find a physical location.

In the above, the Agent is not broadcasting it's Access ID to the world in the same way a Commlink or toaster would in order to be a part of the net. The agent is leaving its data trail for actions it takes on the net while in a node.

Does this all sound reasonable?

Well, in your scenario, you could just follow it at Stpe 2, as you have an active trail you can follow.

The trail would only be "active" as long as the subscription was active. You would have to spend time to get the AID of the subscription. Which makes a race. No?

edit: I'm not even sure you can analyze a subscription, it sounds reasonable though. (nm, you can SR4A 232)

Passive Trace rules are in Unwired.

The closest thing I can find is under Spotting Traces p104, and that doesn't seem to fit.

Hmn.

I thought it was in that section, but apparently they broke it up somewhere else.

Page 65: the access log: A spider can use the information in the access log to Track an intruder through the Matrix (p. 219, SR4 ) even if the intruder’s
icon is no longer in the node.
Combine with spotting traces: you only get to spot a trace if you're in the node when it's started.

The rules in unwired say you can Track by only the Access ID, then point you to the rules in SR4 on how to do it, and the rules in SR4 say you can not use Track in that manner. SR4 says you need an icon with an active connection, or a subscription.

I think I have to assume Unwired is wrong, you can not Track with only an Access ID, because doing so makes the whole "Track from the same node, with a connection up; vulnerable to Redirect, attack, and Track program crashes" not exist. I simply analyze you, get your Access ID, then Track from a different node.

Thats a load of drek. You're confusing Icon with Avatar.

And (232: Trace user): You trace an icon back to its originating node.
(216: A Whole New World): "The Matrix is populated by icons, each a representation of a person, program, file, connection, data stream, control, or anything else with which you can interact"
You can launch a trace attempt against anything, not necessarily an active Avatar of a Persona. Its not guaranteed to work, but you can do it. Also, you misread that bit about subscriptions: You don't need a subscription, but - and note that this is not a mandatory use, but an optional one 0 "You can also use this action to trace a subscription to its other end. For example, you may trace the subscription from a drone to the rigger controlling it."


Anyway, unwired allowing passive traces does make sense. If you disallow it, it means that the data trail is irrelevant. It literally doesn't matter. As soon as someone logs out, you can't touch them forever. You can hack into Zurich Orbital, AAA corps, anyone you like and then 'oh wait, you can't passive trace me or dig up my datatrail that i left because I'm not online anymore! haha i win!'
No, fuck that.
You always leave a data trail, and it can always be used against you if someone bothers to do their legwork.* Spoof your access ID's often.

*note, per unwired 65, backtracking hacks from an IP log/access log often means you're going to have to hack into the originating node to pry into their access logs and find another link and repeat the process. Much like real hacking forensics.

No confusion between icon and avatar.

The confusion is that there are three distinct meanings for track. There is the (1)Track action using the (2)Track program (note my use of capital letters) that is done like phone tracing is done in the movies. You need the person on the line. There needs to be an active connection for the icon you are trying to trace and the icon needs to be in your node. You can also use the Track action with the Track program to also trace a subscription which is just an active connection minus the icon.

Then there is the more mundane (3)tracking which I guess is using Datasearch + Browse to track down a person's data trail when there is no active connection. This is the equivalent of going node to node and asking "Have you seen this ID?"


I very heavily disagree that only knowing a string of 1s and 0s as a hacker's Access ID will let you trace their location directly (Track action with Track program), without a connection, subscription, hacking, or any other step. Even though it is obvious the Dev that wrote Unwired believes that's how it works. That Dev references the main book for the rules on how to track an Access ID with no connection, and the main book pretty much says you trace differently.


If the dev in Unwired is correct, that you can Track action with the Track program on an Access ID without a connection + icon, then the whole description of the Track action in sr4 pg 232 is 100% irrelevant because no one would ever run Track in that manner, it can be noticed, stopped, stalled, or the tracker might even be injured. Tracking is simply to strong if he is correct.

And all this isn't even getting into my original question about independent Agent or AIs and their Access IDs and how they are different then a normal Persona Access ID.