louis@aerospace.aero.org (Louis M. McDonald) (02/16/90)
I have been asked numerious times why someone should purchase and X term when some workstations (especially diskless) are just as cheap). I always give the standard "A workstation requires a sys admin, but an X term does not" I would be interested in hearing from other people as to their thoughts on this issue. Quantitative information would be also appreciated (cost of workstation vs x term) -- Louis McDonald Internet: louis@aerospace.aero.org The Aerospace Corporation 213-336-8914
john@acorn.co.uk (John Bowler) (02/19/90)
In article <66915@aerospace.AERO.ORG> louis@aerospace.aero.org (Louis M. McDonald) writes: >I have been asked numerious times why someone should purchase >and X term when some workstations (especially diskless) are >just as cheap). I always give the standard > >"A workstation requires a sys admin, but an X term does not" > In principle not true, but in practice accurate. A terminal (any terminal, not just X terminals) requires sys admin - but the admin is (mostly) confined to the host (X client) machine(s). In practice this will be a relatively small increment to the administration work required anyway. On the other hand a discless workstation requires looking after in its own right, and there are other consequences:- 1) More machines each with their own environment == more potential for confusion if the environments get out of step. 2) Network security is lower - physical protection of a diskless workstation may be more difficult than that of the file (etc) server machine and it requires (considerably) less expertise to tamper with a diskless workstation than with an X terminal. 3) The interface between a diskless workstation and a file server is far broader than that between an X terminal and an X client server (:-) ie a server of X clients :-):-). This means that more things have to be set up and maintained (in addition to setting up and maintaining the workstation itself). >I would be interested in hearing from other people as to their >thoughts on this issue. Quantitative information would be also >appreciated (cost of workstation vs x term) > Now, as far as I can see the hardware in an X terminal can be exactly the same as that in a diskless workstation (so can the software if you restrict it to bootp type stuff just to load the kernel or X server...) I assert (without proof) that, despite this, X terminal hardware costs less than comparable diskless workstation hardware; that an X terminal with a given (X) performance (measured using any of the widely accepted X benchmarks :-) costs significantly less than a diskless workstation with the same performance. The only problem with proving or disproving these assertions is agreeing on an X benchmark... Why should this be so (if it is)? Because current X servers can get away with significantly less powerful hardware than that required to give adequate performance to a workstation running a multi-processing operating system. John Bowler (jbowler@acorn.co.uk)