Spped of FreeWare FTP-Server under WinXP?

Hi,

I just tested the overall transfere speed of 2 FreeWare FTPds
(Cerberus FTP Server 2.4 BETA 3 and FileZilla Server 0.9.18 beta)
with ftp client Total Commander 6.54a) and 125 MB jpg (holliday) pics
on an 100 MBit LAN ... and the result was so different (*), that I
wonder, if there are more detailed ) speed) tests / experineces around
- but I did not found. So I ask here:


Did anybody ever tested transfere speed of the big 5, Freeware
FTP-Server undder Windows, named
- WarFTPd
- Cerberos FTPd
- GuildFTPd
- zFTPServer Suite
- FileZilla
to see, which Server ist (more or less) the fastest? In my case, the
fastest for many files ... not for one huge file (**).
Is there some commoun test (under Windows) one can Use?


The pure values of my (any test) is not so important, as the depend on
exact conditions ... but the comparea with same conditions of Server
would be. Also I know, that TotalCommander fore sure is not the
fastets FTP client. But I use it mainly, so ...


(*)PS: Here the resoults of my quick test:

125 MB jpeg (385 files in 53 subdirs) , OS: XP to XP

TotalCommander - Cerberos:
1:40 min -> 1.25 MB/sec, to delete them took 15 sec

TolatlCommander - FileZilla Server:
0:40 min -> 3.1 MB/sec, to delete them 10 sec

And no, it is not a cache thing ;-)

(**) If I use a 380 MB Zip File, speed of both was nearly the same:
48 Sec. for FileZilla and 50 Sec. for Cerberous (7.7x MB/sec)

Seems to me there are only a few things that would affect FTP speeds among
different products. To even test this you'd need a carefully controlled LAN
environment to ensure other network traffic/issues aren't impacting the results.
Since FTP protocol is the same regardless of package, the main factors (other
than the network) would have to be speed of the app itself (should make little
difference) and how fast it can read/write to the file system. Again, there
shouldn't be a huge difference in results here which is probably why you don't
see speed results for FTP software. Some apps may send more FTP commands back
to the server which may slow down multiple file transfers or directory listings.

"Olaf Bölling" <(E-Mail Removed)> wrote in message
news:(E-Mail Removed)...
> Hi,
>
> I just tested the overall transfere speed of 2 FreeWare FTPds
> (Cerberus FTP Server 2.4 BETA 3 and FileZilla Server 0.9.18 beta)
> with ftp client Total Commander 6.54a) and 125 MB jpg (holliday) pics
> on an 100 MBit LAN ... and the result was so different (*), that I
> wonder, if there are more detailed ) speed) tests / experineces around
> - but I did not found. So I ask here:
>
>
> Did anybody ever tested transfere speed of the big 5, Freeware
> FTP-Server undder Windows, named
> - WarFTPd
> - Cerberos FTPd
> - GuildFTPd
> - zFTPServer Suite
> - FileZilla
> to see, which Server ist (more or less) the fastest? In my case, the
> fastest for many files ... not for one huge file (**).
> Is there some commoun test (under Windows) one can Use?
>
>
> The pure values of my (any test) is not so important, as the depend on
> exact conditions ... but the comparea with same conditions of Server
> would be. Also I know, that TotalCommander fore sure is not the
> fastets FTP client. But I use it mainly, so ...
>
>
> (*)PS: Here the resoults of my quick test:
>
> 125 MB jpeg (385 files in 53 subdirs) , OS: XP to XP
>
> TotalCommander - Cerberos:
> 1:40 min -> 1.25 MB/sec, to delete them took 15 sec
>
> TolatlCommander - FileZilla Server:
> 0:40 min -> 3.1 MB/sec, to delete them 10 sec
>
> And no, it is not a cache thing ;-)
>
> (**) If I use a 380 MB Zip File, speed of both was nearly the same:
> 48 Sec. for FileZilla and 50 Sec. for Cerberous (7.7x MB/sec)

"Mike Lowery" <(E-Mail Removed)> wrote:

>Seems to me there are only a few things that would affect FTP speeds among
>different products. To even test this you'd need a carefully controlled LAN
>environment to ensure other network traffic/issues aren't impacting the results.
>Since FTP protocol is the same regardless of package, the main factors (other
>than the network) would have to be speed of the app itself (should make little
>difference)

I agree with the assumption, that the app itself is the speedlimiting
factor ... and disagree (or just don't) beleve, that THIS dosent
matter. But belive is on thing, proves are someting else. This is why
I ask for other experiences.

Olaf