Computer technology

[anders lager]

Does anyone know if there's been any articles or forum threads about Traveller computer technology? I'm thinking of how computer networks are thought to work in the T universe. I can imagine planetary networks like the internet, and maybe these can be expanded to in-system networks using maser or some other tightbeam technology, maybe using relay stations to extend the range.

 

[Malenfant]

Ooo, I dunno if the valves and vacuum tubes on those skyscraper-sized Model 9/bis computers can handle a planetary internet... think of the punch cards you'd need too!

 

[epicenter00]

There's plenty of mentions of them, Doctor. Computer technology is one of those running jokes amongst Traveller players that never really gets old.

As Malenfant obliquely pointed out to you, computers are a joke in Traveller. If you look at their size, weight, cost, and power consumption as compared to what they can do, one can only conclude they're still using COLOSSUS, ENIAC, and UNIVAC in the 57th century. Case in point, according to the TNE rulebook (that I happen to have in front of me) a Suleiman-class Scout ship (the infamous flying wedge ship) has three TL15 (oooh ahhh) computers to handle flying. Each of these require 0.55 Megawatts of power each.

0.55 Megawatts. Holy cow.

Obviously, Grandfather pinched off a pocket universe to develop the Commodore PET, a vast computing milestone that humans in the 57th century couldn't even imagine. Grandfather's children never had a chance, still using those Altairs...

In other words, if you're making or fleshing out the Traveller Universe, feel free to plug in whatever computer technology you wish.

 

[TheEngineer]

Hi !

MTs referees sourcebooks provides a little bit more detailed information about computers/technology.

But whats the problem with 0,55 MW ?
Take a look ....
http://www.llnl.gov/asc/computing_resources/bluegenel/configuration.html

Regards,

TE

 

[Jame]

I asked something similar a good while ago.

Linky for the thread, see if there's anything you like.

 

[atpollard]

IMTU I "fixed" computers by deciding that the model numbers represented vastly different processing powers. Model 1 = kiloflops(10^3); Model 2 = megaflops(10^6); Model 3 = gigaflops(10^9); etc.

I also moved the basic model 1 computer from the vacuum tube TL 5 to the microchip TL 7 and changed the program number to the minimum computer that will run it (evade 2 requires a model 2 computer or higher).

But this is just MY solution.

 

[Malenfant]

TE - that's an earth-based supercomputer. You don't need one to run a spaceship (even a jump-capable one). Keep in mind also that the computer you currently have sitting on your desktop today would have taken up at least a room about 10 or 20 years ago.

As it is, Apollo's guidance computer only required 70W of power (see http://www.computerpoweruser.com/editorial/article.asp?article=articles/archive/c0612/28c12/28c12.asp&guid= ). The shuttle apparently has four computers for redundancy - while I haven't found their actual power requirements I'm fairly certainly they're going to be using a few hundred watts at most.

Even though Traveller spacecraft have fusion plants that are pumping out power in the megawatt range, I think the computers are going to be taking the least amount of that - 550,000 W is a ridiculous amount for even a four-computer system that you'd run on a spacecraft.

It all really depends on what the computers are used for, and the technology. If quantum computers get off the ground (apparently one is going to be demonstrated next month) then that will drastically change computer technology. QCs currently can only be used for specific tasks, but if you need to say calculate jump parameters b between point A and point B then that sounds like something a QC is ideal for. It sounds like you'd need much smaller computers (and less power) to do calcaulations with those - a QC desktop would have the same processing power as a massive supercomputer.

 

[anders lager]

Well, the issue of the massive 70's ship's computers has been debated a lot, I know that.
What I meant was how are communication networks thought to function.
OK, we have the X-boat system, but surely some other means of shorter range data communication must exist...or?

 

[anders lager]

By the way, thanks for your comments. This is a really nice forum, with tons of ideas for Travellers.

 

[flykiller]

in a game, unless you're trying to work out some kind of detail sometimes it's better not to specify how something works. I always just talk about "comm links" - you comm link to the port authority, you have a link request, the link breaks, a one-time-use link authorization number. if I must describe anything I talk about it like it's the internet.

 

[The Shaman]

IMTU the ship's computer and the bridge represent everything from control panels to avionics to sensors as part of their combined tonnage, so the large numbers don't faze me.

As far as the original question, an interplanetary internet is still going to have to move data at no more than the speed of light, so lag times may be a significant issue. Tight-beam comm systems for data transfer between a mainworld corporate HQ and its subsidiary on a jovian gas giant seems reasonable, and depending on planetary ownership, you might have public systems for this as well.

IMTU systems with A or B starports and Pop 4+ allow starships to access local nets by comm.

 

[mbrinkhues]

I have always seen the ship computers as something to the old Sicomp-M series (The Boys as my old boss called them). Highly specialised process control systems with an extremly rugged and redundant hardware setup.

This page has something on the M80, the final member of the series but the more common M70 and M76 series are about the same size. Most commonly used in

+ Nuclear power plants (All german/KWU plants use it)
+ Coal fired power plants (Either M or R-series)
+ Mining (All german mines used it till the 2000s)

The actually could work without an environmental control if necessary. That "5-45 degrees celsius" was a recommendation, nothing more.

They where beautiful maschines working at 99.99 and better reliability handling 1000+ A/D inputs per second. Dual box (Two CPU/Data access, 2 Disk towers, 1 switching unit) in mining and IIRC 2x3 (Two mirrorer tripple boxes) in the nuke plants. You could hot plug cards and on a few occasions the Siemens techs soldered wires on a life box.

And the grafik terminal where simply a great design. Basically an 8bit CPU with a dedicated 6845 CRT Controller and 64KB (yes Kilobyte) of memory for the grafics. Quite smart boxes doing some procession on their own and sturdy enough that a grown man could stand on them. Typically stacked (there where two in grafic terminals) and used as a footrest. Basically the size of a SparcStation each.

The pulled their final "hurray" in the mining jobs when the "supperior follow up" system was unabel to do their job for almost 2 years after the sheduled replacement date. God, I miss those boxes. Big, rugged, serviceable with an electronics soldering iron and with handbooks. Ohhhh the handbooks.

 

[TheEngineer]

Hi !

Originally posted by Malenfant:
TE - that's an earth-based supercomputer. You don't need one to run a spaceship (even a jump-capable one). Keep in mind also that the computer you currently have sitting on your desktop today would have taken up at least a room about 10 or 20 years ago.

Well, Mal, I have to partly disagree

Point is, that a Traveller ship computer is as fictional as devices like reactionsless thrusters or jump drives. Its a kind of computer needed to control those starship components.

There is absolutly no way to exactly define in real world units how much volume or how much power such a computer would use, as its about a fictional thing.
So the units given in Traveller IMHO should just be accepted as they are. Its just not possible to argue if they are correct or not.
My intention was to show, that real world supercomputer systems are another league as the tiny boxes regular people know and that those maschines indeed occupy a vast amount of space and need extraordinary power supplies (besides, the 1,5 MW given in the link are only the power usage of the primäry systems, secondary colling etc is not included).
IMHO it was (perhaps just) luck, that there are no real world stats inside the Traveller rule set

So definitions like

Computer, Hand:
Provides services of a supercomputer
(equivalent to Modell/l in computing power), plus serves as a computer terminal when linked to a larger computer (such as on board a ship).
TL Vol Weight Price
11 0.2 liters 0.5 kg Crl 000

are pretty well, aren't they ?

Regards,

TE

 

[veltyen]

550,000 W is a ridiculous amount for even a four-computer system that you'd run on a spacecraft.

But a nice amount of power for a car.

http://en.wikipedia.org/wiki/Pagani_Zonda

OK I admit - even the current Zonda is under 550KW, but close enough for engineering work

From elsewhere:
Based on current plans to meet LHC offline computing requirements, we will need to make available a false floor area of some 2000m2 by 2003-4 and to support an electrical load of 2MW.

Guess what. That's the machine room for a particle accelerator. CERN to be exact.
https://webh12.cern.ch/it-support-mrp/

My intention was to show, that real world supercomputer systems are another league as the tiny boxes regular people know and that those maschines indeed occupy a vast amount of space and need extraordinary power supplies

I've tried and failed before. It is difficult to convince people outside of industry that a desktop does not equal a reliable, well functioning, redundant, survivable system.

Hell, I work with midrange SUN's, and massively parrallelised intel boxes, not even anything in the high performance computing realm. Even with a strong lee-way for outages (some of my systems have a minimum recovery time greater then 24 hours) the hardware adds up.

There is something ... humbling ... about standing next to a UPS bigger then you are.

 

[mbrinkhues]

Nothing beats the dual UPS we used:

+ A 45MW unit
+ A 850MW unit

That's BIG!

 

[epicenter00]

Originally posted by doctor:
Well, the issue of the massive 70's ship's computers has been debated a lot, I know that.
What I meant was how are communication networks thought to function.
OK, we have the X-boat system, but surely some other means of shorter range data communication must exist...or?

Communications between planets is actually pretty spotty. In fact, it's very spotty, unless you're on a Xboat route, which doesn't actually cover every planet. In fact, fewer planets are on the Xboat routes than are off of them.

X-boats are essentially a relay-rider system. They're actually the second fastest way to communicate between stars. The fastest is to be a part of certain parts of Imperial government and you'll get access to the TJ (jump-6) shadow courier fleet. Otherwise, you're stuck hoping to get messages through with mail traveling at like Jump-2 or something.

In-System communications in the 3I are handled primarily by technologies we know of today - radio, laser, maser, and at higher TLs, meson communicators which travel at the speed of light, so there can be significant delays between sending messages from one end of a solar system to the other.

There are no FTL communications methods in the 3I that are widely and commercially used, though certain psions can transmit messages from system to system or across solar systems without any kind of time delay. Certainly the Zhodani would use such communications.

Reading between the lines, I'm under the impression that the 3I actually has rules and protocols on how communications are handled - Imperial High Law probably has sections about what bands and wavelengths are reserved for use by the Imperium so that ships can talk to each other. In the article about how the AI Virus was developed in Challenge, there's a discussion that every ship in the 3I is basically talking to each other all of the time - computer to computer, relaying course information to COACC and other ships and similar information, sending out transponder information and taking it in (and checking bona fides) all without even the crew's knowledge.

I hope that helps.

 

[Border Reiver]

Worth reading Travellers' Digest (issue 8 I think?) article on computers and data systems in Traveller, some interesting stuff on how Library Data works.

Of course it is all out of date but still makes good reading, there is also an nice bit about why on Sylea a shipboard computer needs so much power.

 

[Malenfant]

Right on cue...
http://news.bbc.co.uk/2/hi/technology/6354225.stm

So a single 80-core chip today requires 62W of power. That gives you the same processing power as a 2000 square foot, 500 kilowatt supercomputer with 10,000 Pentium Pro chips could produce 11 years ago. Literally a supercomputer in your desktop, if they ever actually release this thing for commercial use.

 

[Border Reiver]

Parkinson's Law.

 

[TheEngineer]

Hi !

Anyway regarding communication, there is a short chapter (2 pages) in MTs referees companion.
Well it deals less with networks but with in-system communication.
Classical radio is still the workhorse for civilian long range commo. At least in the presented designs nearly every starship is equipped at least with a system wide ranging radio.
Guess above TL 7/8 everything in civilized regions is heavily networked (hard-wiring/satellites).
Long range backbones (e.g. to other in-system planets, farports etc.) could be established via laser comm or even classical radio.
Another breakthru perhaps happens at TL14, when Meson communication is used as a backbone.

Longer ranges might force communication to work more or less asynchronous, with packaged data (voice/image mail etc..), except you want to wait 5+ Minutes between sends and receives...

But I guess nobody want to play Imperial Counterstrike with a ping of 300000 ms.

Most developed system might provide a broadband readonly datastream, containing the most important data/messages for an incoming vessel.
More specific information could be provided on-demand via special channels.
The MT companion book tells about standard channels and channel directories:

For example, upon entering the Regina system, a ship has information about standard channels such as Navy, Scout, Distress, and Hailing It doesn’t know system-specific channels such as Market Prices, Passage Reservations, Entertainment
Once the Directory has been accessed, the ship’s computer knows what channels are available.

Regards,

TE