Computer Hardware, Ratings cap makes no sense Options

[Stormdrake]

Matrix hardware according to the rules is capt at 6 for Response and Signal. While this makes sense for portable units (like comlinks) to me, it falls apart when you get to computers that mega corporations would be using for running arcolgies or (god forbid) guiding a ship to Mars and back. The rules mention that corporations still use wired systems within their complexes to cut down on their vulnerability. So large scale systems still have to exsist some what don't they? If I am reading the rules right (and I could not be god knows, lol) it is very possible for a player character to have hardware and software hanging off their belt that is just as powerful as anything the corps can martial which seems silly considering the corps (realitive) unlimited assets. Coments?

[Ranneko]

I would assume that corporate systems, wired, have no need for signal and will have a response high enough that it is irrelevant.

So the corp's servers will run as many programs as you feel it needs to.

[Elldren]

Considering that the team hacker is unlikely to be dragging a server cluster with him, I wouldn't worry about the response ratings of such systems. Yes, they're going to be higher. How much higher is really irrelevant.

[Brazila]

I have to say the above signature about ninjas made my day.

[Stormdrake]

Ok so if I am saying that signanl and and response are pretty much meaningless for major corporation computers what about the software (Firewall/system)? Soft ware run off of such systems should have higher ratings than anything that can be run by a comlink, I am thinking. The way I am looking at this is that a comlink should allow a hacker to attempt entrence into vastly faster more advanced computers but it is the deckers skill that allows him to offset the descrepencies between his machine and those he or she is breaking into. Witht he current rules though this is not the case. As it sits now the computer system running an arcology and the one sitting on a hackers hip can be of the same speed and ability which just does not make sense. Granted I reliese that the computer running the arcology is supposed to be made up of hundreds if not thousands of nodes which through sheer numbers should defeat any players hacker from owning the system but it still seems silly that such a system can not mount a firewall/system that outstreches anything a simple comlink can do.
Thoughts?

[Elldren]

[Fortune]

[Rotbart van Dain...]

Assume that the given Response of a host only corresponds to the session to that user?

[Dutchy]

Big Corps may have node attributs above the ordinary 6. Everything else would be unrealistic. For a arcology mainframe a 10 for Response should be a good choise. Maybe higher.

[Rotbart van Dain...]

Most certainly much higher, given the fact that hundred of programs run on those systems.

But it makes also sense to limit the available resources per session... which gets us back to an effective response ranging from 1-6.

[Mightyflapjack]

Honestly in 2070, a mainframe computer would probably be the size of a briefcase? Maybe a large suitcase at most. The only bulk would be in power consumption and cooling system, but if we have "cranial" cyberdecks then we obviously have some nice ways to dissapate computer heat without cooking people's brains.

Sure... Bulk storage of data would be heavy, but not too terrible. You could probably fit a few Terapulses ( 1tp = 1,000 gigapulses = 1,000,000 mp) into a storage unit the size of a large toolbox, but it would weight 10-15 kilos.

So I figure you could fit a megacorportation mainframe PC into a decent sized van, maybe a large passenger car with modifications.

--

Sorry if this is a sidetrack.. I am working on my own house rules with new 2070s electronics and the thread caught my eye.

House Rule = 2070s found FTL (Faster then Light) communications. (Tachyons baby!) http://en.wikipedia.org/wiki/Tachyon.

With Tachyons.. Wireless Matrix finally made sense (to me).

[nezumi]

Funny how when computers were first invented, they took up an entire room. Yet now, more than a century and a half after the first difference engine, every respectable business has one or more 'computer rooms'.

They get smaller, so we buy more. I suspect a super computer will simply be like our current blade server setup. It's just multiple computers you can slot into one large object. So they will continue to be large, as everyone and his brother requires their own blade to to run their personal matrix space, telecomm, so on and so forth.

[Stormdrake]

<<But it makes also sense to limit the available resources per session... which gets us back to an effective response ranging from 1-6.>>

For the typical user on the mainframe I can see this. However what I am looking at is a hacker attempting to break into said mainframe. shouldn't the firewall for such a computer considering whats already been said about the hardware of such a machine be easily over a 6? To say nothing of the IC and Agent programs? My real problem is that a character who choses to have a 6 in hacking and spends the appropriate amount of funds at creation can have a system that rivals anything a megacorp can field.

[Rotbart van Dain...]

Not quite - upgrading Hardware or Software is in the Computer Group and is not limited by Skill.

If an Hacker tried to access much more cpu time than an usual session, he would become too obvious, too.

[hobgoblin]

"there is one user online but the workload is enough for 100, i think we have a hacker"

[ef31415]

[Stormdrake]

That makes some sense. However the complaint being thrown back at me is that if a corp computer was running a firewall program of say ten a hacker would have no chance at getting in. Now my reading of the rules say that the hacker gets his skill plus program or logic to roll to hack in. Not sure which, sorry. if this is right then as an extreme example saying the hacker has maxed out the appropriate stats she could roll twelve dice and get into the system, right?

[sapphire_wyvern]

You can't credibly hack a Firewall 10 system on the fly.

Using the slow probing method it might be do-able though.

The problem is not gaining access; it's doing it without triggering an alert in the process...

[hahnsoo]

[wagnern]

A friend of mine had a job working on miniframes (small mainframes). He said that it was about as fast as a 486 computer (this was a few years ago). Now, I asked why these expencive machines were used unstead of a few servers. According to him, the miniframe was slow, but it was the same slow wether five or one hundred people were waiting on it. (the difference between speed and bandwith). Perhaps the super uber mega computers of 2070 are not that much faster than a normal mainframe, but they have the bandwidth to handle many more task at the same time.

(PS the computer guiding a ship to mars would be a relativly slow one. Reliability is key, and a lot of power is not needed. I beleave the Spaceshuttle's main computer is equilivent to a 286 or such.)

[warrior_allanon]

[hobgoblin]

many mainframes are 99% storage these days.

a IBM big iron from what i understand dont have the biggest of cpus, but have dedicated hardrive controller chips that can basicly swamp a gigabit network connection each.

these things are not crays, they a buildt to handle bulk amounts of raw data traffic to and from storage. most likely the real math workhorses will be a cluster of desktop computers wired up and running some specialy modified version of a os over some very special connections. not high capasity ones tho, but low latency ones. ie, they cant handle much data pr time unit, but the data will not take long to get there.

[Kagetenshi]

That's not necessarily true—some computationally intensive tasks just aren't parallelizable to the degree necessary for a cluster. Clusters really are a sucky solution to a lot of problems, but they're a cheap sucky solution, so they're popular.

~J

[hobgoblin]

true, if the result is based on a chain of calculations the only way to go at it is by trowing one fast cpu at it.

still, most of the time you can break a task down i sections, like say brute forcing a encryption key, and spread it among the cpus.

another thing is that clusters are very flexible. you can segment them up at times when you have many small tasks or bunch em up when there is one big task.

you can expand the power by basicly adding more off the chelf computers to the network.

thing is basicly that they are very flexible for their cost. but they are not the ultimate solution to all tasks. sometimes a specialized computer is a better solution.