RAID is a very useful way of impressing your geek acquaintances, with the useful byproduct of enhanced data storage. It means your data is preserved even if one or two of your disks fail, or it means your data is thingyed faster than theirs, or both.
However, the most failure prone part of any computer
system is the user^H^H^H^H printer and so it seems to me far more logical to propose RAID principles for printing purposes. i.e. RAIP
The user purchases a number of identical printers. They are connected to the raip thingy which then connects to the computer. When you press the big "PRINT" button, instead of your document being sent to the printer, it is sent to the raip whatsit, which divides it between the printers according to how the toggle switches on the raip device are set.
e.g. RAIP 1 consists of print mirroring. Print jobs are sent identically to two or more printers, thereby producing a "mirrored set" of printouts. Thus, any filing, sending or other use can be serviced by any printout in the set.
RAIP 4 consists of page-level striping with dedicated parity. The main advantage of RAIP 4 over RAIP 2 and 3 is page parallelism: in RAIP 2 and 3, a single page output requires printing using the whole group of printers, while in RAIP 4 each page print operation is not have to spread across all printers. As a result, the text on one page of the document is printed on a single piece of paper, and does not have to be reassembled using scissors and glue.
RAIP 6 consists of page-level striping with double distributed parity. Double parity provides fault tolerance up to two failed printers. This makes larger RAIP groups more practical. RAIP 6 requires a minimum of four printers. As with RAIP 5, a single paper jam results in reduced performance of the entire array until the jammed piece of paper has been pulled out of the innards of the printer. The larger the print job, the more important it becomes to choose RAIP 6 instead of RAIP 5. RAIP 10 also minimizes these problems.