If there is any gain in speed, it is probably in the logic of the cards
rather than in any cabling. The basic physics of a transmission
line/cable at 1 GHz (or greater if you want clean data with fast rise
times) mean that I doubt that these systems can really work reliably in
the real world at 1 GBit/s, even if the computers at either end could
cope. To be fair, I havent seen CAT6 cable - how thick is it, and is it
shielded?
Richard
Greg Twyford wrote:
Peter Machell wrote:
Greg, there's a lot more to it than this. Gigabit LAN is bottlenecked by
the speed of the PCI bus and in real terms is around twice as fast as
100baseT if you are lucky.
Peter,
4 seconds to open a patient's notes, versus 25+ seconds, sounds more
like 6 to 1 to me. The systems I've set up with onboard gigabit NICS
and switch record 55 MBPS, using Sandra as the measuring tool, against
9-10 MBPS for 100 mbit NICs and switch over the same cables.
I know that on a properly set up PCI-X bus, their rated speed is more
like 90+ MBPS, but we are talking about making life easier for GPs
now, not some other time. And yes, I expect MD3 to be much faster, as
it is client-server and will send less traffic over the LAN.
However, like TS, the penalty is the need for a more robust server,
and possible serious additional licensing costs, especially for larger
practices where the MSDE won't cut it. For smaller practices TS is not
cheaper, and if implemented in larger practices, properly, needs a
very good server and reliable, not moribund, clients with good
graphics, as Horst has suggested. Also, the advantages of TS diminish
further if you are using a client-server application in the first place.
Greg
_______________________________________________
Gpcg_talk mailing list
[email protected]
http://ozdocit.org/cgi-bin/mailman/listinfo/gpcg_talk
