<!DOCTYPE html PUBLIC "-//W3C//DTD HTML 4.01 Transitional//EN"> <html> <head> <meta content="text/html;charset=ISO-8859-1" http-equiv="Content-Type"> </head> <body bgcolor="#ffffff" text="#000000"> Pretty much all of the fiber ones, which is about all I use, are that way. One good one that I've found works well with Linux is Amer.com's C1000SX. It is available here: <a class="moz-txt-link-freetext" href="http://www.amer.com/catalogue/ac1000sx.html">http://www.amer.com/catalogue/ac1000sx.html</a> Another one is the Intel Pro/1000SX card, which is 64bit/66MHz. These have been around for a long time, and I have used them in many a GNU/Linux box. Just do a search on <a class="moz-txt-link-abbreviated" href="http://www.pricewatch.com">www.pricewatch.com</a>, using the terms "intel gigabit fiber", and I find them reasonably priced. As for copper cards, those typically are integrated into the motherboard nowadays, even on client box models. That's a question that the motherboard manufacturer should answer--on which bus, and at what bit/speed rate, is their integrated 10/100/1000 network interface(s)? Tyan and MSI both are known to put the copper Gig-E interface either on the PCI-32 bus (lower-end mobos) or, in Tyan's case, also on the PCI-X bux (higher end, co$t$ more). If it's the latter, you're in good shape. If it's the former, but you're running your other Gig-E cards on the PCI-X bus, then other than IDE hard disk contention (also on the PCI-32 bus, but not usually constant), you probably will be fine. Of course, I'd be looking at SATA or SCSI hardware raid on PCI-X if the budget allows for it. --TP <div class="moz-signature">_______________________________ Do you GNU!? <a href="http://www.gnu.org/">Microsoft Free since 2003</a>--the ultimate antivirus protection! </div> Petre Scheie wrote: <blockquote cite="mid45C36A05.1070808@maltzen.net" type="cite"> Terrell Prudé Jr. wrote: <blockquote type="cite">Robert Arkiletian wrote: <blockquote type="cite">On 1/31/07, Robert Arkiletian <a class="moz-txt-link-rfc2396E" href="mailto:robark@gmail.com"><robark@gmail.com></a> wrote: <blockquote type="cite">On 1/31/07, Petre Scheie <a class="moz-txt-link-rfc2396E" href="mailto:petre@maltzen.net"><petre@maltzen.net></a> wrote: <blockquote type="cite">Terrell Prudé Jr. wrote: <blockquote type="cite">Robert Arkiletian wrote: <blockquote type="cite">On 1/29/07, Joseph Bishay <a class="moz-txt-link-rfc2396E" href="mailto:joseph.bishay@gmail.com"><joseph.bishay@gmail.com></a> wrote: <blockquote type="cite">Hello, I hope you are doing well. Thank you all for the comprehensive reply! Once I started reading your email, I realized that probably the </blockquote> </blockquote> </blockquote> </blockquote> best <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">way to proceed was to work with the idea of NIC Bonding or port trunking. I have a surplus of Gigabit cards so I could put 3 in a server (reading online I found that more than 3 wasn't going to </blockquote> </blockquote> </blockquote> </blockquote> give <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">enough of an improvement due to the PCI bus limitations -- can </blockquote> </blockquote> </blockquote> </blockquote> anyone <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">validate this?) and then send all 3 of those to the switch. I </blockquote> </blockquote> </blockquote> </blockquote> could <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">then bond 3 ports from that switch to the next one (we'll probably have 2 x48 gigabit switches for the whole building -- still </blockquote> </blockquote> </blockquote> </blockquote> counting <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">the number of ports/computers required) so as to deal with the bandwidth. The cost of some of those fiber <-> copper converts </blockquote> </blockquote> </blockquote> </blockquote> look <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">rather daunting. I would VERY MUCH prefer to use only 1 server for the entire </blockquote> </blockquote> </blockquote> </blockquote> building <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">-- I am still very much a novice at this and the complexities of setting up multiple servers or splitting into application & </blockquote> </blockquote> </blockquote> </blockquote> /home with <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">LAPD sounds rather daunting. </blockquote> If your still set on one server also have a look at this <a class="moz-txt-link-freetext" href="http://k12ltsp.org/mediawiki/index.php/Technical:Subnetting">http://k12ltsp.org/mediawiki/index.php/Technical:Subnetting</a> Instead of port trunking I think this would be a better idea. Especially if you are going to have 2 48 port switches that </blockquote> </blockquote> </blockquote> could be <blockquote type="cite"> <blockquote type="cite"> <blockquote type="cite">on different gigabit linked subnets. </blockquote> Hmm...I hadn't thought of that particular application </blockquote> </blockquote> myself--addressing <blockquote type="cite"> <blockquote type="cite">bandwidth bottlenecks--but you're right, that sure would do it! </blockquote> </blockquote> That <blockquote type="cite"> <blockquote type="cite">never even occurred to me...thanks! --TP </blockquote> I recall reading somewhere that three gigabit cards is probably the </blockquote> max that the PCI bus <blockquote type="cite">can handle. Can anyone confirm or deny this? </blockquote> No. A gigabit card is 1 Gibabit/s (that's 1 billion bits per second). Each byte is 8 bits. So it maxs out at 125MB/s. A simple PCI bus can handle 133MB/s max. So 1 gigabit ethernet card can saturate a PCI bus </blockquote> Correction: PCI 2.2 spec is 32 bits at 66Mhz which equals 266MB/s. So 2 gigabit nics should be able to saturate it. The original PCI bus was 32bits at 33Mhz which is 133MB/s. </blockquote> True, but if your PCI bus is 64-bits at 66MHz (i. e. PCI-X), then you're fine, as you then have 532MB/s. I've always been sure to buy 64-bit, 66MHz NIC's for this reason. Same with RAID cards; PCI-X whenever possible. </blockquote> What brand of 64-bit NIC are you buying for this purpose? Where do you get them? Petre _______________________________________________ K12OSN mailing list <a class="moz-txt-link-abbreviated" href="mailto:K12OSN@redhat.com">K12OSN@redhat.com</a> <a class="moz-txt-link-freetext" href="https://www.redhat.com/mailman/listinfo/k12osn">https://www.redhat.com/mailman/listinfo/k12osn</a> For more info see <a class="moz-txt-link-rfc2396E" href="http://www.k12os.org"><http://www.k12os.org></a> </blockquote> </body> </html>