 More about Agents and IC |
  |
 May 19 2006, 06:20 AM
Post
#26
|
|
|
Beetle Eater  Group: Dumpshocked Posts: 4,797 Joined: 3-June 02 From: Oblivion City Member No.: 2,826  |
In SR3 this was called a "Virtual Machine"; I imagine SR4 would simply use mulitple clone nodes because they are so damn cheap.
|
 |
 
|
|
May 19 2006, 08:15 AM
Post
#27
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
fluff text something about traffic and security and whatever |
||
|
|
|
||
|
May 19 2006, 04:01 PM
Post
#28
|
|
|
Ain Soph Aur Group: Dumpshocked Posts: 3,477 Joined: 26-February 02 From: Montreal, Canada Member No.: 600 |
Hmm, picking up on a post here.. idea brewing:
Ok, the Node can run Analyse, right? So, and I'm not looking at the rules right now as I write tis, what if everytime you loaded an Agent, the Node could detect it. Or simply detect that a user is loading too many agents on the Node. I mean, if the Agent doesn't have Stealth anyway, the Node will automatically detect, right? If that happens, the Node would automatically either launch IC to crash those Agents, simply as a precautionary method, or go full blown into Alert and look for a hacker and kill all those Agents. Anyway, my point is, it shouldn't be too hard for a Node to detect a DoS attack is under way, and do somethin about it. |
|
|
|
|
May 19 2006, 04:56 PM
Post
#29
|
|
|
Moving Target Group: Members Posts: 914 Joined: 26-August 05 From: Louisville, KY (Well, Memphis, IN technically but you won't know where that is.) Member No.: 7,626 |
Actually, if the Agent is running on the node the node should be aware of it. It may not recognize it as an Agent per se but it will recognize that an application has been loaded that is taking resources.
I suppose one solution to prevent DDOS would be for the node to instigate a block on loading new applications. I mean, the node has to download the excecutable code and allocate resources at some point. |
|
|
|
|
May 19 2006, 05:01 PM
Post
#30
|
|
|
Target Group: Members Posts: 6 Joined: 14-May 06 From: Waltham, MA Member No.: 8,561 |
Why would Agents and IC cost so damn much if they were intended to be copy-able?
|
|
|
|
|
May 19 2006, 05:22 PM
Post
#31
|
|||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
So a DoS attack in SR4 is simply overloading the node so it's reponse drops considerably. So what if all the legit users go into the node and all pop out their Browse agents? That's what the issue is. It makes no sense that it would only take a handful of users to reduce the resouces of a node. My solution is that agents (per the text) don't reduce the node, but IC (per the text) does. This would means a 2006 DoS style attack cant happen unless you overloaded a node with IC somehow. |
||
|
|
|
||
|
May 19 2006, 05:22 PM
Post
#32
|
|
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
@Epoch:
Why does Adobe Suite 2.0 cost so much? Is it copyable? @Grinder: Every Corp would only run agents in your world, not IC . . . But agents and IC are, per RAW, exactly the same thing, anyways . . . |
|
|
|
|
May 19 2006, 05:30 PM
Post
#33
|
|||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
From the what I stated it earlier, Personas run agents and Systems run IC. IC is a subset of an agent. Ask youself why they needed to make a distinction between the two if they are identical? The differences are slight, but they do exist. |
||
|
|
|
||
|
May 19 2006, 05:49 PM
Post
#34
|
|||
|
Moving Target Group: Members Posts: 371 Joined: 10-January 06 From: Regina Member No.: 8,145 |
You are on the right track I think. So a Hacker can load up to System x 2 Agents onto a target Node. However, since each of those is considered an independent "user" (and they would have to be to run on their own) then the Node gets a chance to detect each of them just like the Hacker. Now, the Threshold to detect an intruder is the Stealth program, so if the Agent isn't running Stealth then they are automatically detected. The rules clearly state that your Agent is tied to your Persona, whether you are online or not. So, if a Trace is run on the Agent they find your Commlink. I think I'm starting to see the counter-point here. Sure your Hacker can bring a ton of Agents, but for every Agent he puts on the system he is giving the Host another chance to detect him, essentially. Shoot, if a PC Hacker wants to spend 24 Complex Actions to load 12 Agents on to the Host, then let him. If the Node/IC is rolling well they could have the Trace completed before they're even done. Wow. Now I see the danger. Bringing 12 Agents in with you on a 'run is like the Sammie kicking down the door of the corp building and opening up with an EX-EX equipped LMG. Not exactly quiet. |
||
|
|
|
||
|
May 19 2006, 06:20 PM
Post
#35
|
|||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
Unless you're feeling lucky and build Agents that can hack in themselves. Plus, unless you have Admin access, there's probably no reason *any* user should be dumping more than 2-3 agents at a time in a node, that'd trigger paranoid security to start looking at you more closely. If you probed your taget system instead of brute hacking, I think you could squeeze in more explot/stealth agents with the probed exploit but the system still gets a roll to detect them. |
||
|
|
|
||
|
May 19 2006, 06:35 PM
Post
#36
|
|||||||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
They do not make a distinction.
(emphasis by me) Please be aware of th fact that by RAW, there is no disticntion between matrix nodes and normal nodes or any devices, and there is no difference between agents and IC. It is perfectly valid to claim that there must be a distinction to make the rules consistent (I myself amkind of Mr. Houoserule), but do not try to justify this by SR4 rules. It simply is not there and only confuses people. |
||||||
|
|
|
||||||
|
May 19 2006, 06:43 PM
Post
#37
|
|||||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Only when your comlink is running the agent. Not anymore after it is uploaded. And definately not if you are really offline.
This is out of question. Bringing even one agent in is a danger to stealth. Everybody should be aware of this. I think we are talking about the danger of an extremely efficent method to force the node to shut down. That is the only thing what lots of agents do. Nobody in his right mind would question the fact that once you load tons of agents you wont be stealthy any more. |
||||
|
|
|
||||
|
May 19 2006, 07:00 PM
Post
#38
|
|
|
panda! Group: Members Posts: 10,331 Joined: 8-March 02 From: north of central europe Member No.: 2,242 |
if the options are having valuable data "stolen" (more correctly copyed, but anyone know that the value of data comes from its rarity) and taking the node offline, i would take the "node offline" option each time...
|
|
|
|
|
May 19 2006, 08:30 PM
Post
#39
|
|||||
|
Moving Target Group: Members Posts: 371 Joined: 10-January 06 From: Regina Member No.: 8,145 |
Well even if it is not quite RAW I am going to rule that all of your Agents contain your "signature" and thus can be used to Trace your Commlink. If you're offline, the Trace will fail, but they should have enough information (Commcode) to determine when you are back online to start the Trace. |
||||
|
|
|
||||
|
May 19 2006, 08:38 PM
Post
#40
|
|||||||||||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
Here's my statement from before:
I still don't understand why you would need to define something twice. If Agents = IC and IC = Agents, then you'd only need to define it once. So there must be some difference. If it's been printed twice I'd argue there is some reason for it, else why ever specify one from the other? |
||||||||||
|
|
|
||||||||||
|
May 19 2006, 08:56 PM
Post
#41
|
|||||
|
Ain Soph Aur Group: Dumpshocked Posts: 3,477 Joined: 26-February 02 From: Montreal, Canada Member No.: 600 |
Yes, I like this too. |
||||
|
|
|
||||
|
May 19 2006, 11:09 PM
Post
#42
|
|
|
Runner Group: Members Posts: 2,556 Joined: 26-February 02 From: Seattle Member No.: 98 |
IC programs are agents specifically designed to respond to intruders. This means that they work like agents, move like agents, are combated like agents... they are, in fact, agents (in the same way that a Porche is a specific kind of car).
However, when you talk about Agents, you aren't necessarily talking about IC (that Agent could be running Edit, and constantly filtering data you feed it... not an Intrusion Countermeasure). This makes it pretty obvious to me that IC are simply a subtype of Agent. |
|
|
|
|
May 20 2006, 03:10 AM
Post
#43
|
|||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Because humans wrote the rules, and several of them? |
||
|
|
|
||
|
May 20 2006, 03:16 AM
Post
#44
|
|||||||
|
Running Target Group: Members Posts: 1,498 Joined: 4-August 05 From: ADL Member No.: 7,534 |
Why? Your commcode is only used for commcalls. You can have a totally different ID while loggin in into the matrix. Your matrix ID is like an IP number. A commcode could be associated to a certain IP number or to no number at all (like you might have to log on to the commserver to recieve commcalls and you are recognised by login and password). To give a modern day example: There is only one way to find out which IP number started a certain task on a server. You can find out which user started the task and which IP was associated to the user at that time. If you delete that log entry, there is nothing left. You would have to invent a very queer IP/ID system to justify your ruling. |
||||||
|
|
|
||||||
|
May 22 2006, 04:04 AM
Post
#45
|
|||||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
Tounge and cheek aside my point is valid. It's been suggested I reading too much into the rules and maybe that's true, I just get the impression that exact wording is key considering some of the questions we are all raising. |
||||
|
|
|
||||
|
May 22 2006, 02:06 PM
Post
#46
|
|||
|
Mr. Johnson Group: Dumpshocked Posts: 3,148 Joined: 27-February 06 From: UCAS Member No.: 8,314 |
I'm sitting in a computer lab right now. We have, among other machines, a workstation and a server that are identical other than the purpose to which they are put. The person wrapped in duct tape watching you complete your shadowrun is just as much a person as the one with the assault rifle responding to the security alert and shooting at you. And yet one is a wage slave and the other a security guard. That's the difference. |
||
|
|
|
||
|
May 22 2006, 02:55 PM
Post
#47
|
|
|
panda! Group: Members Posts: 10,331 Joined: 8-March 02 From: north of central europe Member No.: 2,242 |
basicly, the IC uses the normal agent rules but behave diffrently in that its a kind of agent that is launched by the node automaticaly as a reaction to detecting a intrusion. some may even be roaming as guards.
ie, they are system launched rather then user launched, and keep their IC tag mostly as a way of telling them apart ;) hmm, i wonder how unwired will deal with tricks like cluster/party IC and constructs. or for that matter cascading. that is if they even get ported over (i hope they do). |
|
|
|
|
May 22 2006, 03:02 PM
Post
#48
|
|
|
Runner Group: Members Posts: 2,925 Joined: 26-February 02 Member No.: 948 |
As Hobgoblin said. The IC is the agent that the system itself can launch if it realizes that it has been infiltrated. The system can choose the appropriate programs to load the IC with and can continue to upload IC until someone shuts down the system.
Agents can be the first line of defense and as soon as an Agent bites the dust the OS can decide to launch IC in order to keep up the defenses. This makes it clear that the OS itself has some limited decision making abilities and/or that the owner can specify what conditions should apply for the OS to activate IC. So the main difference could be: Agents: Can only be loaded by user but can roam the matrix IC: Can be activated by OS but remains inside the node |
|
|
|
|
May 22 2006, 05:12 PM
Post
#49
|
|||||
|
Shooting Target Group: Members Posts: 1,754 Joined: 9-July 04 From: Modesto, CA Member No.: 6,465 |
I understand your analogy, but I based on how I am reading the Agent/IC section I don't see how they are identical. If they where truely identical and if they shared all things, then at the end of desribing Agents/IC add "Agents/IC count towards the their controlling nodes program count (see Pg.212)" but it doesn't. It spells that part out specifically under Intrustion Countermeasures. They've said IC inherits Agent properties, but didn't say Agent inherits IC properties. They are similar, but not the same. It's all up for interpretation and based on the language presented in SR4.227-228, this is how I see it. |
||||
|
|
|
||||
|
May 22 2006, 05:36 PM
Post
#50
|
|||
|
Mr. Johnson Group: Dumpshocked Posts: 3,148 Joined: 27-February 06 From: UCAS Member No.: 8,314 |
If that's the case, though, their design is rather inefficient. But I'll bite: what, if not purpose, makes them different? They don't even have different costs. |
||
|
|
|
||
|
|
Lo-Fi Version | Time is now: 3rd December 2025 - 04:47 AM |
Topps, Inc has sole ownership of the names, logo, artwork, marks, photographs, sounds, audio, video and/or any proprietary material used in connection with the game Shadowrun. Topps, Inc has granted permission to the Dumpshock Forums to use such names, logos, artwork, marks and/or any proprietary materials for promotional and informational purposes on its website but does not endorse, and is not affiliated with the Dumpshock Forums in any official capacity whatsoever.