Professional Web Applications Themes

10baseT ethernet - bottleneck or not? - Mac Networking

I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT onboard ethernet performs so badly in comparison to the two newer and wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line via a Netgear DG824M modem/router. The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP machine with Netgear wireless PCI card - are happily downloading at over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it bursts to approx 150kb/sec for the first few seconds of any download. I wondered if the onboard ethernet was ...

  1. #1

    Default 10baseT ethernet - bottleneck or not?

    I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    onboard ethernet performs so badly in comparison to the two newer and
    wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    via a Netgear DG824M modem/router.

    The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    machine with Netgear wireless PCI card - are happily downloading at
    over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    bursts to approx 150kb/sec for the first few seconds of any download.

    I wondered if the onboard ethernet was the problem since I'm the only
    one using wires and a specialist networking techie at Apple UK support
    confirmed that 10baseT would not be able to handle the full 2Mbps. He
    said I should divide 10 by 8 to get the max throughput, ie 1.25Mbps.

    I reported this to my ISP and he said that the Apple techie was wrong
    and that I should see the full 2Mbps via 10baseT. His full reply was:
    >The network may be 10Mbps, but it's well-known that Ethernet , being a
    >CSMA/CD type network, will not achieve anything like that in practice,
    >although 2Mbps does seem extremely low. The figure I had in mind was
    >something like 50%-70% of the available bandwidth, but one reference I
    >found quoted 40%:
    >
    >The CSMA/CD article wrote:
    >"Performance of CSMA / CD
    >
    >"It is simple to calculate the performance of a CSMA/CD network where
    >only one node attempts to transmit at any time. In this case, the node
    >may saturate the network and near 100% utilisation of the network may
    >be achieved, providing almost 10 Mbps of throughput on a 10 Mbps LAN.
    >
    >"However, when two or more nodes attempt to transmit at the same time,
    >the performance of Ethernet is less predictable (and not covered by
    >this course). The fall in utilisation and throughput occurs because
    >some bandwidth is wasted by collisions and back-off delays. In
    >practice, a busy shared 10 Mbps Ethernet network will typically supply
    >2-4 Mbps of throughput to the nodes connected to it.
    >
    >"As the level of utilisation of the network increases, particularly if
    >there are many nodes competing to share the bandwidth, an overload
    >condition may occur. In this case, the throughput of Ethernet LANs
    >reduces very considerably, and much of the capacity is wasted by the
    >CSMA/CD algorithm, and very little is available for sending useful
    >data. This is the reason why a shared Ethernet LAN should not connect
    >more than 1024 computers. Many Engineers use a threshold of 40%
    >Utilisation to determine if a LAN is overloaded. A LAN with a higher
    >utilisation will observe a high collision rate, and likely a very
    >variable transmission time (due to back off). Separating the LAN in to
    >two or more collision domains using bridges or switches would likely
    >provide a significant benefit (assuming appropriate positioning of the
    >bridges or switches).
    >
    >"Shared networks may also be constructed using Fast Ethernet, operating
    >at 100 Mbps. Since fast Ethernet always uses fibre or twisted pair, a
    >hub or switch is always required."
    >
    >
    >A home network with a handful of nodes would not normally be considered
    >a "busy network", so it does raise the question of what protocols are
    >in use on the network? Is a lot of unnecessary traffic being generated?
    >
    >
    >On the face of it, it does seem unlikely that a 10Mbps Ethernet should
    >not be able to keep pace with your 2Mbps/256kbps ADSL connection.
    Can anyone here please clarify this for me. Who is right? TIA.

    Stan
    Stan The Man Guest

  2. #2

    Default Re: 10baseT ethernet - bottleneck or not?


    "Stan The Man" <machomac.com> wrote in message
    news:220920030008177794%machomac.com...
    > I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    > onboard ethernet performs so badly in comparison to the two newer and
    > wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    > via a Netgear DG824M modem/router.
    >
    > The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    > machine with Netgear wireless PCI card - are happily downloading at
    > over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    > bursts to approx 150kb/sec for the first few seconds of any download.
    >
    > I wondered if the onboard ethernet was the problem since I'm the only
    > one using wires and a specialist networking techie at Apple UK support
    > confirmed that 10baseT would not be able to handle the full 2Mbps. He
    > said I should divide 10 by 8 to get the max throughput, ie 1.25Mbps.
    I'd say its more likely the mac itself thats the problem. The car may well
    be able to manage 10MBps, but the graphics updating, file downloading and
    file saving to harddisk may not be able to cope. I'm running a 10base2 lan
    here with an LC475 included. The LC475 is far slower on the network than
    anything else, but I've always put it down to overall hardware. I would
    imagine that if you have an ethernet card, it should be designed to use the
    full network 'speed' - perhaps you can find some specs for it on the apple
    support site?
    >
    > I reported this to my ISP and he said that the Apple techie was wrong
    > and that I should see the full 2Mbps via 10baseT. His full reply was:
    I think you were speaking to someone who doesn't deal with old equipment and
    just assumes everything will work at the same speed because it has the same
    type of card. Everything he says below is right to my knowledge, but I think
    irrelevant to your issue.
    >
    > >The network may be 10Mbps, but it's well-known that Ethernet , being a
    > >CSMA/CD type network, will not achieve anything like that in practice,
    > >although 2Mbps does seem extremely low. The figure I had in mind was
    > >something like 50%-70% of the available bandwidth, but one reference I
    > >found quoted 40%:
    > >
    > >The CSMA/CD article wrote:
    > >"Performance of CSMA / CD
    How old is this article? I get the feeling its almost talking about my type
    of cabling, 10base2, where everything's along one wire.
    > >
    <cut>
    > >
    > >
    > >A home network with a handful of nodes would not normally be considered
    > >a "busy network", so it does raise the question of what protocols are
    > >in use on the network? Is a lot of unnecessary traffic being generated?
    That does raise a question - are you running Appletalk over the ethernet
    aswell? I have no idea but that might slow it down (?) Especially on an
    older mac.
    > >
    > >
    > >On the face of it, it does seem unlikely that a 10Mbps Ethernet should
    > >not be able to keep pace with your 2Mbps/256kbps ADSL connection.
    >
    > Can anyone here please clarify this for me. Who is right? TIA.
    >
    > Stan
    Sorry I can't be of more help!
    matt

    --
    matthew fullerton
    LC475/MacOS 7.5.5 and LCIII/System 7.1
    This message was posted using Outlook Express 6 (sorry)
    [email]greatbigblackholehotmail.com[/email] is a real address, but don't expect anything
    (SPAM or otherwise) to get read



    ---
    Outgoing mail is certified Virus Free.
    Checked by AVG anti-virus system ([url]http://www.grisoft.com[/url]).
    Version: 6.0.520 / Virus Database: 318 - Release Date: 18/09/2003


    matt Guest

  3. #3

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <bkmfmf$a03$1pump1.york.ac.uk>, matt
    <greatbigblackholehotmail.com> wrote:
    >"Stan The Man" <machomac.com> wrote in message
    >news:220920030008177794%machomac.com...
    >> I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    >> onboard ethernet performs so badly in comparison to the two newer and
    >> wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    >> via a Netgear DG824M modem/router.
    >>
    >> The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    >> machine with Netgear wireless PCI card - are happily downloading at
    >> over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    >> bursts to approx 150kb/sec for the first few seconds of any download.
    >>
    >> I wondered if the onboard ethernet was the problem since I'm the only
    >> one using wires and a specialist networking techie at Apple UK support
    >> confirmed that 10baseT would not be able to handle the full 2Mbps. He
    >> said I should divide 10 by 8 to get the max throughput, ie 1.25Mbps.
    >
    >I'd say its more likely the mac itself thats the problem. The car may well
    >be able to manage 10MBps, but the graphics updating, file downloading and
    >file saving to harddisk may not be able to cope. I'm running a 10base2 lan
    >here with an LC475 included. The LC475 is far slower on the network than
    >anything else, but I've always put it down to overall hardware. I would
    >imagine that if you have an ethernet card, it should be designed to use the
    >full network 'speed' - perhaps you can find some specs for it on the apple
    >support site?
    Hardware limitations are a possibility but I would doubt it since
    almost everything has been upgraded (cpu to G4, internal HD to big
    7200rpm IBM drive, RAM to 740MB, FW card, USB card, etc) and I use this
    machine as my daily workhorse in preference to any of the newer
    machines around the place. It seems pretty snappy all round - even in
    comparison to my G4 TiBook running Os X and 9) - but of course it still
    does have the original motherboard and scsi internal drive buses. I may
    try a 100baseT ethernet PCI card to see if that improves anything.
    >>
    >> I reported this to my ISP and he said that the Apple techie was wrong
    >> and that I should see the full 2Mbps via 10baseT. His full reply was:
    >
    >I think you were speaking to someone who doesn't deal with old equipment and
    >just assumes everything will work at the same speed because it has the same
    >type of card. Everything he says below is right to my knowledge, but I think
    >irrelevant to your issue.
    >
    >>
    >> >The network may be 10Mbps, but it's well-known that Ethernet , being a
    >> >CSMA/CD type network, will not achieve anything like that in practice,
    >> >although 2Mbps does seem extremely low. The figure I had in mind was
    >> >something like 50%-70% of the available bandwidth, but one reference I
    >> >found quoted 40%:
    >> >
    >> >The CSMA/CD article wrote:
    >> >"Performance of CSMA / CD
    >
    >How old is this article? I get the feeling its almost talking about my type
    >of cabling, 10base2, where everything's along one wire.
    No idea, but you could be right.
    >> >
    ><cut>
    >> >
    >> >
    >> >A home network with a handful of nodes would not normally be considered
    >> >a "busy network", so it does raise the question of what protocols are
    >> >in use on the network? Is a lot of unnecessary traffic being generated?
    >
    >That does raise a question - are you running Appletalk over the ethernet
    >aswell? I have no idea but that might slow it down (?) Especially on an
    >older mac.
    I am using Appletalk - set to Ethernet - to communicate with the ADSL
    modem/router. I also need AT on in order to run postscript printer
    drivers, afaik. And since there is no wireless facility on the 7500, I
    don't know how to fileshare from it without using AT. I would love to
    kill off AT but I don't think it's possible.
    >
    >> >
    >> >
    >> >On the face of it, it does seem unlikely that a 10Mbps Ethernet should
    >> >not be able to keep pace with your 2Mbps/256kbps ADSL connection.
    >>
    >> Can anyone here please clarify this for me. Who is right? TIA.
    >>
    >> Stan
    >
    >Sorry I can't be of more help!
    >matt
    Stan
    Stan The Man Guest

  4. #4

    Default Re: 10baseT ethernet - bottleneck or not?


    "Stan The Man" <machomac.com> wrote in message
    news:220920031202051048%machomac.com...
    > In article <bkmfmf$a03$1pump1.york.ac.uk>, matt
    > <greatbigblackholehotmail.com> wrote:
    >
    > >"Stan The Man" <machomac.com> wrote in message
    > >news:220920030008177794%machomac.com...
    > >> I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    > >> onboard ethernet performs so badly in comparison to the two newer and
    > >> wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    > >> via a Netgear DG824M modem/router.
    > >>
    > >> The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    > >> machine with Netgear wireless PCI card - are happily downloading at
    > >> over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    > >> bursts to approx 150kb/sec for the first few seconds of any download.
    > >>
    > >> I wondered if the onboard ethernet was the problem since I'm the only
    > >> one using wires and a specialist networking techie at Apple UK support
    > >> confirmed that 10baseT would not be able to handle the full 2Mbps. He
    > >> said I should divide 10 by 8 to get the max throughput, ie 1.25Mbps.
    > >
    > >I'd say its more likely the mac itself thats the problem. The car may
    well
    > >be able to manage 10MBps, but the graphics updating, file downloading and
    > >file saving to harddisk may not be able to cope. I'm running a 10base2
    lan
    > >here with an LC475 included. The LC475 is far slower on the network than
    > >anything else, but I've always put it down to overall hardware. I would
    > >imagine that if you have an ethernet card, it should be designed to use
    the
    > >full network 'speed' - perhaps you can find some specs for it on the
    apple
    > >support site?
    >
    > Hardware limitations are a possibility but I would doubt it since
    > almost everything has been upgraded (cpu to G4, internal HD to big
    > 7200rpm IBM drive, RAM to 740MB, FW card, USB card, etc) and I use this
    > machine as my daily workhorse in preference to any of the newer
    > machines around the place. It seems pretty snappy all round - even in
    > comparison to my G4 TiBook running Os X and 9) - but of course it still
    > does have the original motherboard and scsi internal drive buses. I may
    > try a 100baseT ethernet PCI card to see if that improves anything.
    >
    This just goes to show my lack of knowledge about (new) macs! I thought it
    was an older machine. Your problem is indeed very strange :s

    matt

    > >>
    > >> I reported this to my ISP and he said that the Apple techie was wrong
    > >> and that I should see the full 2Mbps via 10baseT. His full reply was:
    > >
    > >I think you were speaking to someone who doesn't deal with old equipment
    and
    > >just assumes everything will work at the same speed because it has the
    same
    > >type of card. Everything he says below is right to my knowledge, but I
    think
    > >irrelevant to your issue.
    > >
    > >>
    > >> >The network may be 10Mbps, but it's well-known that Ethernet , being a
    > >> >CSMA/CD type network, will not achieve anything like that in practice,
    > >> >although 2Mbps does seem extremely low. The figure I had in mind was
    > >> >something like 50%-70% of the available bandwidth, but one reference I
    > >> >found quoted 40%:
    > >> >
    > >> >The CSMA/CD article wrote:
    > >> >"Performance of CSMA / CD
    > >
    > >How old is this article? I get the feeling its almost talking about my
    type
    > >of cabling, 10base2, where everything's along one wire.
    >
    > No idea, but you could be right.
    > >> >
    > ><cut>
    > >> >
    > >> >
    > >> >A home network with a handful of nodes would not normally be
    considered
    > >> >a "busy network", so it does raise the question of what protocols are
    > >> >in use on the network? Is a lot of unnecessary traffic being
    generated?
    > >
    > >That does raise a question - are you running Appletalk over the ethernet
    > >aswell? I have no idea but that might slow it down (?) Especially on an
    > >older mac.
    >
    > I am using Appletalk - set to Ethernet - to communicate with the ADSL
    > modem/router. I also need AT on in order to run postscript printer
    > drivers, afaik. And since there is no wireless facility on the 7500, I
    > don't know how to fileshare from it without using AT. I would love to
    > kill off AT but I don't think it's possible.
    > >
    > >> >
    > >> >
    > >> >On the face of it, it does seem unlikely that a 10Mbps Ethernet should
    > >> >not be able to keep pace with your 2Mbps/256kbps ADSL connection.
    > >>
    > >> Can anyone here please clarify this for me. Who is right? TIA.
    > >>
    > >> Stan
    > >
    > >Sorry I can't be of more help!
    > >matt
    >
    > Stan


    --
    matthew fullerton
    LC475/MacOS 7.5.5 and LCIII/System 7.1
    This message was posted using Outlook Express 6 (sorry)
    [email]greatbigblackholehotmail.com[/email] is a real address, but don't expect anything
    (SPAM or otherwise) to get read



    ---
    Outgoing mail is certified Virus Free.
    Checked by AVG anti-virus system ([url]http://www.grisoft.com[/url]).
    Version: 6.0.520 / Virus Database: 318 - Release Date: 18/09/2003


    matt Guest

  5. #5

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <220920031202051048%machomac.com>,
    Stan The Man <machomac.com> wrote:
    > Hardware limitations are a possibility but I would doubt it since
    > almost everything has been upgraded (cpu to G4, internal HD to big
    > 7200rpm IBM drive, RAM to 740MB, FW card, USB card, etc) and I use this
    > machine as my daily workhorse in preference to any of the newer
    > machines around the place. It seems pretty snappy all round - even in
    > comparison to my G4 TiBook running Os X and 9) - but of course it still
    > does have the original motherboard and scsi internal drive buses. I may
    > try a 100baseT ethernet PCI card to see if that improves anything.
    Setting aside any confusion about bits and bytes, which may be the
    explanation for the weird response from the Apple engineer, you ought to
    be able to do better...

    I just tried sending files with Timbuktu to my old back-up server which
    is a stock 7200 running System 8.1 using onboard ethernet and achieved
    550kB/s over Appletalk and 760kB/s over TCP/IP pushing a 6MB pdf file
    (all numbers bytes not bits here), so a pretty impressive go at
    flattening the 10Mb/s / 1.25MB/s ethernet into the 7200.

    Same game to 7500 with G3, System 9.2.2, fast internal SCSI disk on
    internal bus and fast ethernet gives both Appletalk and TCP/IP at around
    1.4 MB/s. My general impression is that here you're getting near the
    limit of what the 7500 and stock SCSI can do.

    My recollection is that downloads on the 7500 with IE typically peaked
    in the 200-300 kB/s range.

    So even if your original figures were bytes rather than bits things look
    sick - if they really were bits then everything is poor and that might
    be suggestive of the answer...

    See if you can find an old copy of MacTCP watcher as that gave useful
    information on the state of incoming duplicate packets and outgoing
    retransmits. Try pinging your other computers to see if you're OK to
    them, even if the rest of the world is a bit foggy.

    Dave
    David Sankey Guest

  6. #6

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <220920030008177794%machomac.com>,
    Stan The Man <machomac.com> wrote:
    > I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    > onboard ethernet performs so badly in comparison to the two newer and
    > wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    > via a Netgear DG824M modem/router.
    >
    > The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    > machine with Netgear wireless PCI card - are happily downloading at
    > over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    > bursts to approx 150kb/sec for the first few seconds of any download.
    >
    > I wondered if the onboard ethernet was the problem since I'm the only
    > one using wires and a specialist networking techie at Apple UK support
    > confirmed that 10baseT would not be able to handle the full 2Mbps. He
    > said I should divide 10 by 8 to get the max throughput, ie 1.25Mbps.
    >
    > I reported this to my ISP and he said that the Apple techie was wrong
    > and that I should see the full 2Mbps via 10baseT. His full reply was:
    >
    > >The network may be 10Mbps, but it's well-known that Ethernet , being a
    > >CSMA/CD type network, will not achieve anything like that in practice,
    > >although 2Mbps does seem extremely low. The figure I had in mind was
    > >something like 50%-70% of the available bandwidth, but one reference I
    > >found quoted 40%:
    > >
    > >The CSMA/CD article wrote:
    > >"Performance of CSMA / CD
    > >
    > >"It is simple to calculate the performance of a CSMA/CD network where
    > >only one node attempts to transmit at any time. In this case, the node
    > >may saturate the network and near 100% utilisation of the network may
    > >be achieved, providing almost 10 Mbps of throughput on a 10 Mbps LAN.
    > >
    > >"However, when two or more nodes attempt to transmit at the same time,
    > >the performance of Ethernet is less predictable (and not covered by
    > >this course). The fall in utilisation and throughput occurs because
    > >some bandwidth is wasted by collisions and back-off delays. In
    > >practice, a busy shared 10 Mbps Ethernet network will typically supply
    > >2-4 Mbps of throughput to the nodes connected to it.
    > >
    > >"As the level of utilisation of the network increases, particularly if
    > >there are many nodes competing to share the bandwidth, an overload
    > >condition may occur. In this case, the throughput of Ethernet LANs
    > >reduces very considerably, and much of the capacity is wasted by the
    > >CSMA/CD algorithm, and very little is available for sending useful
    > >data. This is the reason why a shared Ethernet LAN should not connect
    > >more than 1024 computers. Many Engineers use a threshold of 40%
    > >Utilisation to determine if a LAN is overloaded. A LAN with a higher
    > >utilisation will observe a high collision rate, and likely a very
    > >variable transmission time (due to back off). Separating the LAN in to
    > >two or more collision domains using bridges or switches would likely
    > >provide a significant benefit (assuming appropriate positioning of the
    > >bridges or switches).
    > >
    > >"Shared networks may also be constructed using Fast Ethernet, operating
    > >at 100 Mbps. Since fast Ethernet always uses fibre or twisted pair, a
    > >hub or switch is always required."
    > >
    > >
    > >A home network with a handful of nodes would not normally be considered
    > >a "busy network", so it does raise the question of what protocols are
    > >in use on the network? Is a lot of unnecessary traffic being generated?
    > >
    > >
    > >On the face of it, it does seem unlikely that a 10Mbps Ethernet should
    > >not be able to keep pace with your 2Mbps/256kbps ADSL connection.
    >
    > Can anyone here please clarify this for me. Who is right? TIA.
    >
    > Stan
    I have an old 7600 upgraded with a 500 MHz G3 runing MacOS 9.1 connected
    by 10 MHz Ethernet to an Asante Router, which in turn connects to
    Comcast cable modem service. The 7600 can max out the cable modem at
    around ~200 MByte/sec, as can my other newer Macs running 10.2.6. The
    old 7600 also got ATA disks, a PCI graphics card, and 1 GB of memory a
    few years ago.
    Alan Charlesworth Guest

  7. #7

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <220920030008177794%machomac.com>,
    Stan The Man <machomac.com> wrote:
    > I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    > onboard ethernet performs so badly in comparison to the two newer and
    > wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    > via a Netgear DG824M modem/router.
    >
    > The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    > machine with Netgear wireless PCI card - are happily downloading at
    > over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    > bursts to approx 150kb/sec for the first few seconds of any download.
    >
    It is correct that Ethernet real maximum throughput is *roughly* half
    what one might expect from the 10 megabits per second implied by
    10baseT. (I didn't include that part of your quote of the Apple guy.)

    But...more likely your problem is Apple's. The built-in Ethernet port
    on Macs of roughly the 7500 vintage tended to be broken out of the box,
    and produce very slow speeds. The built in interface on my 7300 (not
    the same thing, but similar vintage) has been broken from the start. I
    could improve it somewhat by shuffling memory around, but it was never
    workable.

    I simply replaced it with a 10baseT card (later replaced again with a
    10/100 card--cheaper of course than the 10baseT card was). These things
    are dirt cheap, although how much longer you'll be able to find drivers
    for them for ancient Macs is another question.

    --John

    --
    Email to above address discarded by provider's server. Don't bother sending.
    John Baxter Guest

  8. #8

    Default Re: 10baseT ethernet - bottleneck or not?

    John Baxter <news.collectivizescandaroon.com> claimed:
    >
    > But...more likely your problem is Apple's. The built-in Ethernet port
    > on Macs of roughly the 7500 vintage tended to be broken out of the box,
    > and produce very slow speeds. The built in interface on my 7300 (not
    > the same thing, but similar vintage) has been broken from the start. I
    > could improve it somewhat by shuffling memory around, but it was never
    > workable.
    >
    I don't think it was "broken" per se. It was intended to run on 10Mb/s
    networks and it did that reasonably well. When people started plugging
    in 10/100 devices it fell apart because it couldn't negotiate speed or
    duplicity...
    Peter KERR Guest

  9. #9

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <user-0DFAC1.10144325092003scream.auckland.ac.nz>,
    Peter KERR <userhost.domain> wrote:
    > John Baxter <news.collectivizescandaroon.com> claimed:
    > >
    > > But...more likely your problem is Apple's. The built-in Ethernet port
    > > on Macs of roughly the 7500 vintage tended to be broken out of the box,
    > > and produce very slow speeds. The built in interface on my 7300 (not
    > > the same thing, but similar vintage) has been broken from the start. I
    > > could improve it somewhat by shuffling memory around, but it was never
    > > workable.
    > >
    >
    > I don't think it was "broken" per se. It was intended to run on 10Mb/s
    > networks and it did that reasonably well. When people started plugging
    > in 10/100 devices it fell apart because it couldn't negotiate speed or
    > duplicity...
    >
    I have a 7300/200 whose stock motherboard 10 BT Ethenet works fine
    plugged into a 10/100 switch, which in turn plugs into a four-port
    Asante router, which connects to a cable modem. The 7300 acesses the
    cable modem and does Appletalk over IP based file transfers to my newer
    100 BT and Airport Macs just fine.

    The 7300 does have a 500 MHz G3 CPU upgrade, an IDE disk upgrade via a
    PCI card, and a gig of memory. It runs MacOS 9.1. My recollection was
    that Appletalk networking peformed much better when I got the G3 CPU
    upgrade several of years ago.
    Alan Charlesworth Guest

  10. #10

    Default Re: 10baseT ethernet - bottleneck or not?

    In article
    <news.collectivize-A7EBE2.17525423092003corp.supernews.com>, John
    Baxter <news.collectivizescandaroon.com> wrote:
    >In article <220920030008177794%machomac.com>,
    > Stan The Man <machomac.com> wrote:
    >
    >> I've been trying to figure out why my vintage 7500/Os9.1 with 10baseT
    >> onboard ethernet performs so badly in comparison to the two newer and
    >> wireless machines on the same LAN, all hooked up to a 2Mbps ADSL line
    >> via a Netgear DG824M modem/router.
    >>
    >> The wireless machines - a G4 PB (Os 10.2.6 + Airport card) and a WinXP
    >> machine with Netgear wireless PCI card - are happily downloading at
    >> over 200kb/sec but my 7500 can only manage around 50kb/sec, albeit it
    >> bursts to approx 150kb/sec for the first few seconds of any download.
    >>
    >
    >It is correct that Ethernet real maximum throughput is *roughly* half
    >what one might expect from the 10 megabits per second implied by
    >10baseT. (I didn't include that part of your quote of the Apple guy.)
    >
    >But...more likely your problem is Apple's. The built-in Ethernet port
    >on Macs of roughly the 7500 vintage tended to be broken out of the box,
    >and produce very slow speeds. The built in interface on my 7300 (not
    >the same thing, but similar vintage) has been broken from the start. I
    >could improve it somewhat by shuffling memory around, but it was never
    >workable.
    >
    >I simply replaced it with a 10baseT card (later replaced again with a
    >10/100 card--cheaper of course than the 10baseT card was). These things
    >are dirt cheap, although how much longer you'll be able to find drivers
    >for them for ancient Macs is another question.
    >
    I have now installed a Kingston 10/100 card and it's working but it
    hasn't improved my throughput. Downloads are maxing out at 50k/sec and
    uploads are around 12k/sec, measured over several days.

    Simon
    Stan The Man Guest

  11. #11

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <101020031243156850%manpr100.com>,
    Stan The Man <manpr100.com> wrote:
    > I have now installed a Kingston 10/100 card and it's working but it
    > hasn't improved my throughput. Downloads are maxing out at 50k/sec and
    > uploads are around 12k/sec, measured over several days.
    I have a Kingston card sitting around unused because it didn't work any
    better than the 10Base-T. It seemed to be a driver issue. I had a
    later version of the driver from their website (which was later pulled),
    but while the data transfer would go fast, I'd get corruption. With the
    latest version they now have on their website, I'd get slow transfers.

    I replaced it with an Apple 10/100 card. No problems under either OS 9
    or OS X and it transfers quickly on my Beige G3.

    Greg B.

    --
    Actual e-mail address is gbuchner and I'm located at mn.rr.com
    Greg Buchner Guest

  12. #12

    Default Re: 10baseT ethernet - bottleneck or not?

    In article <101020031243156850%com>, Stan The Man
    <com> wrote:
     
    > >
    > >It is correct that Ethernet real maximum throughput is *roughly* half
    > >what one might expect from the 10 megabits per second implied by
    > >10baseT. (I didn't include that part of your quote of the Apple guy.)
    > >
    > >But...more likely your problem is Apple's. The built-in Ethernet port
    > >on Macs of roughly the 7500 vintage tended to be broken out of the box,
    > >and produce very slow speeds. The built in interface on my 7300 (not
    > >the same thing, but similar vintage) has been broken from the start. I
    > >could improve it somewhat by shuffling memory around, but it was never
    > >workable.
    > >
    > >I simply replaced it with a 10baseT card (later replaced again with a
    > >10/100 card--cheaper of course than the 10baseT card was). These things
    > >are dirt cheap, although how much longer you'll be able to find drivers
    > >for them for ancient Macs is another question.
    > >[/ref]
    > I have now installed a Kingston 10/100 card and it's working but it
    > hasn't improved my throughput. Downloads are maxing out at 50k/sec and
    > uploads are around 12k/sec, measured over several days.
    >
    > Simon[/ref]

    I routinely shuffle files between an 8500, a 7600 and various VMS and Unix
    systems, using Fetch 3.03. Pretty much all the Mac - non-Mac connections
    give me hundreds of kb/sec, up to 800...900 between the 8500 & the Unix
    box next to it. Only the Mac-Mac connections seem to be much slower; in
    line with your 50kb/s.

    Uli
    Uli Guest

  13. Moderated Post

    Default Re: 10baseT ethernet - bottleneck or not?

    Removed by Administrator
    Clark Guest
    Moderated Post

  14. #14

    Default Re: 10baseT ethernet - bottleneck or not?

    Stan The Man <com> wrote:
     

    While other posters have said it is the hardware, it is not. It is your
    software. My 8500 which is roughly the same vintage as your 7600 under
    OS 9.1 is roughly 30-40% _slower_ than the *same* machine running
    Linux/PPC; this is from personal experience, as the machine is
    dual-boot. If I need any large downloads, I would switch to Linux/PPC,
    download, and then copy the file to a shared partition. The difference
    in speed far outweighed the time of rebooting to either OS.

    It is an issue with OS9 itself. The newer OS X built on top of *nix no
    doubt has some more efficient/faster code and handling for the Ethernet,
    so operates faster.
     
    >
    > Can anyone here please clarify this for me. Who is right? TIA.
    >
    > Stan[/ref]


    --
    tendim
    remove all capitals from eMail address for replies
    tendim Guest

Similar Threads

  1. Odd performance bottleneck
    By Sandy.Pittendrigh@gmail.com in forum MySQL
    Replies: 1
    Last Post: July 13th, 02:21 PM
  2. laptop bottleneck
    By toberttobert webforumsuser@macromedia.com in forum Macromedia Director Basics
    Replies: 2
    Last Post: February 9th, 08:40 AM
  3. PROBLEM: Notebook w/PCMCIA ethernet & docking station ethernet / Multiple NICs / Only use one at a time
    By ---==[Quasar]==--- in forum Linux Setup, Configuration & Administration
    Replies: 0
    Last Post: September 19th, 04:07 PM
  4. Outbound HTTP Connection Bottleneck
    By Peter Burke in forum ASP.NET Web Services
    Replies: 1
    Last Post: July 10th, 06:34 AM
  5. Ethernet on AIX 1.3
    By Christian Bauer - News in forum AIX
    Replies: 0
    Last Post: July 7th, 08:52 PM

Bookmarks

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •  

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139