Round 3? Kinda wish this topic would stay in its own thread until we can hash it out once and for all eh.
QUOTE
1) Agents are expensive. Especially good ones.
2) A Copied Agent keeps the Access ID of the original. This means that only one of that Agent can connect to any one node at a time. Nodes will not accept multiple connections from the same ID
3) Changing that ID takes atleast 2 weeks if I remember right.
This AND subscription limits. You can only connect so many users before response degredation anyway. Counting legit accounts, necessary connections, slaved nodes, etc. - there is only so much room left over for agents to connect remotely.
The only way to bypass this is with a Botnet (all agents grouped as one), which is very limited command wise, and easy to prevent once the spider's or hacker's are "clued-in" to whats happening. Botnets are usually limited because they are often all copies with the same Access ID, so they really only help with mass probes and DDOS attacks.
Even saying a massive botnet with unique Access ID's that are logged in as one subscription means they can be spoofed as one. I can easily turn that bot army against the node they're attempting to defend.
QUOTE
OK, what we are basically talking about is a BotNet. After all, using a BotNet is really the only way you can control such a vast array of Agents. Thankfully BotNets are talked about in Unwired. Lets have a looksit shall we?
He's all over it.
QUOTE
crizh because so far the "counterarguement" isn't worth responding to, especially when you have yet to answer my simple question which I shall repeat once again.
Wait what? You don't agree, accuse others of being closed minded, but note yourself it isn't worth discussing because you're just so sure your interpertation is the right one? Bad form. What simple question? If you can't see the cost involved, the poor command function, the redundancy as a defense, and are ignoring subscription limits, then I agree, no one can answer your "simple question" to your satisfaction anyway, as stated in the first part of that sentence.
QUOTE
IF Agents worked the way that you claim they do then why doesn't everyone load up as many Agents as they can and set them loose on the Matrix like raging locusts? In the "good 'ol days" this was called the Agent Smith Army and was found to cause far more problems than it solved.
Not as individual agents anyway per the subscription limit and response degredation rules. Unless were talking about botnets (see above). This is taking for granted they each have a unique access ID to begin with.
This is no different than allowing a single rigger to command a horde of drones, possible, but expensive and ultimately easily spoofed or defended against. They cannot just subscribe 1000 drones individually to one commlink (unless they group them all and they all act as one). There are limits to how much can connect to a single node. Bear in mind this analogy is here because agents are referred to as Matrix drones.
QUOTE
The corps literally have more money than God, and the manpower necessary to make the code changes to have their army of Agents, so needing to buy or code seperate Agents is hardly a rebutal.
Oh, then they can all just retire, or maybe all the wage slaves get the luxury lifestyle for free. There is no such thing as department budgets in your world, so no one can argue against your mentality of the infinite. If this is your reasoning, then how about, because GM's should know better than to give corporations unlimited wealth. I'd love to imagine the board meeting where they discuss matrix security and try to get approval for the funding.
QUOTE
And what is complete nonsense is the idea that the corps pay full price on their software, how exactly do you think they are able to program Agents for sale to the masses in the first place?
Lets not forget the
HARDWARE that has to run all those agents anyway. To run remote agents, you're actually paying the limitations cost twice. If I load an agent on my commlink, I pay the cost in processing limit. If I run it remote, I'm paying for separate hardware (if I truly want to "own" it anyway and not have it patched or discovered and removed), that hardware is paying a processing limit (meaning no 200 agents on one node), and i'm paying the subscription limit to command each. This is not even considering how many remote connections the node being defended can support above and beyond what it has allocated to personnel logged in or other needed devices slaved to the master.
Now i'm at work (shame on me) and don't have any books for crunch, but if that's the sole and only reason you reject agents accessing a node without l
oading onto it first, you're just not looking hard enough for why mass agents are specifically limited and excluded by the very rules that allow remote connections.
Edit* Just had to add, under the interpertation that an agent must load onto any nodes it accesses, how do you handle when an IC traces a user back to his home node and engages them in cybercombat there.
Do you now load the offending agent and all his programs onto your player's commlink, calculate the processing load for both the agent and the player, and then initiatie the slowest (if even possible without the link crashing) cybercombat in history?