Re: [Nolug] LAN speed confusion

FTP can be used to measure speeds, but one must be aware of the huge
difference between
1Gb worth of small files (say 4Kb to 300Kb) vs. 1Gb of single file.

I guess the chatter I meant is called TCP overhead.

P

On Sat, Jul 14, 2012 at 1:39 AM, John Souvestre <johns@sstar.com> wrote:

> Hello Petri.****
>
> ** **
>
> What chatter is there other than the normal TCP overhead? Indeed, short
> of dedicated tools, FTP is what is generally used to measure speed.****
>
> ** **
>
> John
>
> John Souvestre - New Orleans LA - (504) 454-0899****
>
> ** **
>
> *From:* owner-nolug@stoney.kellynet.org [mailto:
> owner-nolug@stoney.kellynet.org] *On Behalf Of *Petri Laihonen
> *Sent:* Friday, July 13, 2012 10:40 pm
>
> *To:* nolug@nolug.org
> *Subject:* Re: [Nolug] LAN speed confusion****
>
> ** **
>
> Ron,
>
> Yes. FTP is probably one of the worst protocols AFAIK.
> There is so much back and forth chatter that it consumes a lot of
> bandwidth.
>
> This is most prominent when transferring few hundreds (or thousands)
> smaller files vs. same files inside one tar file. I'm sure most protocols
> have similar problems.
>
> P****
>
> On Fri, Jul 13, 2012 at 6:17 PM, Ron Johnson <ron.l.johnson@cox.net>
> wrote:****
>
> Is ftp really a wasteful protocol?
>
> (Though I do remember that file xfer speeds took a tumble when moving from
> BBS/Zmodem to TCPIP/ftp.)****
>
>
>
> On 07/13/2012 06:06 PM, Petri Laihonen wrote:****
>
> My theory is wasteful protocols. (If such terminology even exists)
>
> I have similar observation. Mine maxed out at 12MB as well, though the
> speed dropped to 7MB after I converted from single drive on the media
> server to Raid6. (Local read/write speeds are now higher, yet transfers
> over the network are slower)
>
> I don't trust interwebs speed tests.
>
> P
>
>
> On Fri, Jul 13, 2012 at 5:26 PM, Ron Johnson <ron.l.johnson@cox.net>
> wrote:****
>
> I've got a bunch of devices at home connected to a Fast Ethernet switch.
> Among them are a Netgear switch (aku), my Linux desktop (haggis) and my
> wife's Linux desktop (peanutbutter). So, I should get a theoretical max of
> 12MBps in half-duplex mode and 25MBps in full-duplex mode.
>
> That's presumably why internet speed tests on both haggis and peanutbutter
> show peaks up to 24MBps (close enough to my paid-for max of 25MBps).
>
> Why, then, do xfers (NFS, scp & plain old ftp) between haggis and
> peanutbutter max out just under 12MBps?
>
> Ron****
>
>
> --
> "There is usually only a limited amount of damage that can be
> done by dull or stupid people. For creating a truly monumental
> disaster, you need people with high IQs."
> Thomas Sowell
>
>
> ___________________
> Nolug mailing list
> nolug@nolug.org****
>
> ** **
>

___________________
Nolug mailing list
nolug@nolug.org
Received on 07/14/12