On page 104 of Unwired there's a section on using a second node (or more in a chain) as a proxy server, but their benefit is described in two ways.

First, all communications sent from the user appear to come from the proxy node, so to trace back to the true source one would need to trace back to the proxy and then examine the access log to find out who's been playing silly buggers, assuming the log hasn't been edited in the meantime.

Second, the user's Response is reduced by 1 for each proxy used but in return all traces against him have the threshold increased by 4 for each.

So what's the difference between using a proxy the first way and using it the second way?
If there is no difference, why would anyone go the second route which offers weaker protection at a price?

They actually are two end results of using a Proxy, one positive and one negative. The Trace will hit the Proxy and stop. If you are using a Proxy Sertver (or Multiple ones), you suffer a degradation of Response (-1 per Proxy Server). However, do not be confused. Proxy Servers are not the same thing as a System, comprised of Individual nodes in a Series, with a gateway seperating each Node. They are NOT the same thing.

Oh I'm not confused on that score.

So, are you saying that both apply simultaneously?

I.e. I set up a proxy and use it while screwing with your node at -1 Response. You try to trace me using the details I forgot to wipe from your access log. You get +4 to your threshold and when you finally reach it your trace leads you to my proxy server. The proxy server happens to be toaster number 4 at the Savoy so you figure this probably isn't the original node. You hack in and check the access log but this time I did remember to wipe off my fingerprints so you give up hope of ever finding me.

Is that about right?

One thing to think about is that there are two ways to do a trace.

There is the "active" trace, using the trace program on a a persona or agent icon, and trying to maintain contact with it until the required hits are collected. But there is also the "passive" trace, going over the various logs and such and backtracking the target that way.

The first sentence could be related to the passive trace, while the second is related to a active trace.

Btw, as i think about it the active trace is something of a oddball holdover from earlier SR editions. If you want to know where someone is located you could just as well hack their node and read off the GPS data of their comlink. But then SR4 matrix rules are a hodgepodge of old and new, where the old has more in common with 80s phreaking than 90s internet (in that each LTG/RTG is a phone switch that needed to be interrogated by the software to figure out if the call was local or rerouted).

Hrmf, now that i think about it perhaps i read to much into the whole accessid thing.

It may well be that said id is more like the id number of a mobile phone sim card than a ip address, and so the SR4 matrix still operates as a "dialup" in the broad sense.

Since they don’t detail it out too well, here are my thoughts on how I think it should work.

Let’s say the hacker sets up 2 proxy servers before she heads in to hack “The Big Bad”. She gets into “TBB” and pokes around setting off a silent alarm.

The system begins tracing her. Because she set up 2 proxy servers “TBB” is looking at a threshold of 18 (10+4+4). Now during this trace if the hacker looks for and finds that there is an active trace she can redirect it, or log off and spoof her access ID and kill the trace entirely. Once it reaches this threshold “TBB” has found the first proxy.

For the sake of argument “TBB” has a standard operating procedure of “hack the target node and gain access to its GPS location, then report this to security.” So the spider hacks the first proxy to gain said info. While the spider is in the node he determines that this is a proxy and begins the trace again. This time the spider has a threshold of 14 (10+4 for only 1 proxy).

Now the whole process starts again. The biggest difference now for the spider is that the hacker can no longer detect or redirect the trace since it is starting in a different node.

Just my 2¥ on the subject.

-D

Yes, that is about right.

Hmm, not convinced. There's no mechanical difference between tracking a datatrail from an icon and tracking a datatrail from the access log so I don't consider drawing a distinction for the purposes of arbitrarily clarifying a separate issue very satisfactory.

This doesn't work for me either – why would it be harder to trace back to the proxy server than it would be if the proxy server had instead been the originating node?

Actually I don't think it is, but I do think I've got it now. It now seems to me that the distinction hinges critically on the reference to "messages" in the second paragraph and "connection" in the third.

That is, sending a communication or command through a proxy makes it look as though it originated at the proxy. You're not logging on to anything or using a subscription so a Response penalty would neither make sense nor have any meaningful impact so that's why it wasn't mentioned in the second paragraph. Therefore, tracing a "message" sent through a proxy will lead them to the proxy.

Whereas, routing your connection through a proxy you are in real-time two-way communication with the target node and (reading between the lines now) the proxy can't treat this like billions and billions of little messages as above because that would be too arduous. Hence, your datatrail can be traced back through a proxy to your originating node but in exchange for a drop in Response the extended test's threshold is increased by 4.

Anyone who thought this was bleeding obvious – thanks for your patience.
Anyone wishing to contest this – I'm happy to discuss further.

Hacking is different than just sending a Data Request (Messages)... Hacking Requires that Subscription you are talking about. Tracing a Hack will also lead them to the Proxy. To go further, they have to delve into the Proxy to root out your data trail from there (which is why the Threshold increases by +4 per proxy used). I do think that the Degradation to Response is a bit harsh (how many times do you see a movie or tv show where the Uber-Hacker routes through 20+ proxies prior to the penetration of a system, or communication with the protagonists - Yes I know, Dramatic License), but I can live with it... Proxies extend the trace time, which is good for the Hacker, especially since Trace is so darn fast anyways.



No Worries...

You make a good point.

Looks like it's time for house rules again, since the RAW doesn't detail out enough information to be useful. *sigh*

It's funny there are combat examples to cover most every possible combat situation, but examples for anything else that might be useful are nearly impossible to find in the books.

So tracing a hacker with an active subscription each proxy sould add the +4 to the threshold and will result in the trace locating the originating node. However sending a single data request (message) through a proxy will only lead a trace to that proxy and the search will have to stop... Is that about right on how you guys see it?

-D

Combat has stayed largely the same over the edition, as has magic (SR4 may have had the largest shift there since voodoo introduced possession).

Matrix however has had 3 complete changes in 4 editions. 1-2 started out with "dungeon" maps, then in late 2 VR2.0 introduced pure numeric hosts. This carried over into SR3 mostly unchanged. SR4 not only introduced yet another wholescale change in the stats, but also wireless and merged matrix and rigging.