[Info-vax] Slow FTP transfer Linux -> VMS

[Bill Gunshannon]

In article <176uZD2KcidF-pn2-IhyX12f74x5U at rikki.tavi.co.uk>,
	"Bob Eager" <rde42 at spamcop.net> writes:
> On Mon, 31 Aug 2009 12:09:34 UTC, billg999 at cs.uofs.edu (Bill Gunshannon)
> wrote:
> 
>> In article <176uZD2KcidF-pn2-RVRggx6Lnwx0 at rikki.tavi.co.uk>,
>> 	"Bob Eager" <rde42 at spamcop.net> writes:
>> > On Mon, 31 Aug 2009 00:20:14 UTC, billg999 at cs.uofs.edu (Bill Gunshannon)
>> > wrote:
>> > 
>> >> In article <1q-dnVwNZJjJfAfXnZ2dnUVZ_radnZ2d at earthlink.com>,
>> >> 	"Michael D. Ober" <obermd. at .alum.mit.edu.nospam.> writes:
>> >> > "Bill Gunshannon" <billg999 at cs.uofs.edu> wrote in message 
>> >> > news:7fre14F2mj5o9U4 at mid.individual.net...
>> >> >> In article <4a9780ef$0$196$e4fe514c at news.xs4all.nl>,
>> >> >> Robin Schipper <"thespriteman <remove_this> writes:
>> >> >>> Hi,
>> >> >>>
>> >> >>> when i try to put a big file (lets say 100 Mb) from a Linux system trugh
>> >> >>> FTP on a VMS system, the max speed is 1,5 MB/s
>> >> >>> when i try to put the same file from another VMS system (on the same
>> >> >>> network) to the VMS box the speed is about 10MB/s (it is a 100Mb network)
>> >> >>>
>> >> >>> does anyone know why the linux box doesn't speed up,
>> >> >>
>> >> >> Because it's linux.
>> >> >>
>> >> >>>                                                       and how to solve 
>> >> >>> it?
>> >> >>
>> >> >> Don't use linux.
>> >> >>
>> >> >> bill
>> >> >>
>> >> >> -- 
>> >> >> Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
>> >> >> billg999 at cs.scranton.edu |  and a sheep voting on what's for dinner.
>> >> >> University of Scranton   |
>> >> >> Scranton, Pennsylvania   |         #include <std.disclaimer.h>
>> >> >>
>> >> > 
>> >> > 
>> >> > Ahh - finally someone who responds with the same BS as when a Windows user 
>> >> > posted similar FTP problems recently.  FTP to one VMS box was slow but to 
>> >> > another VMS box it was fast.  The response then was don't use Windows - use 
>> >> > any other OS.  This sounds like there is a misconfiguration on the slow VMS 
>> >> > box.
>> >> 
>> >> Unless, of course, one has actually looked at Linux and knows what its
>> >> shortcomings are.  One of the worst being the TCP/IP implementation.
>> > 
>> > Evidence? URL?
>> > 
>> > Not that I use Linux...
>> 
>> What evidence do you want?  The Linux IP Stack was written from scratch
>> while dozens of existing usable IP stacks existed.  It performs poorer
>> than any other (anyone who has read even this group for any length of
>> time would have already seen other complaints about poor performance
>> between Linux and other OSes.  If you don't beleive that, benchmark
>> it yourself.  They broke LPD (even MS got that right!!).  They broke NFS.
>> How much evidence do you need?
> 
> Sorry, that's not evidence - it's hearsay. 

Maybe for you, but it's first hand for me.  :-)

>                                             Complaints in this group may 
> well be biased - and even technically inaccurate. Proper, peer reviewed 
> benchmarks?

I have never seen a "Proper, peer reviewed benchmark" in this group and
I doubt I ever will.  What I see is technical experts offering their
expert opinion.  Others are free to take it or leave it.

> 
> Writing it from scratch doesn't mean it's worse. It may be...and you may
> be right. But I'd just like to see some evidence.

NIH Syndrome is highly unlikely to result in a better product when it
causes the developers to ignore more than 20 years of research.  Remember,
while people like to blame the AT&T lawsuit for the rise of Linux and
the failure of BSD the TCPIP code from BSD was never involved in that
lawsuit because none of it came from AT&T.

> 
> (the stack I'm using was written some 30 years ago, although obviously 
> updated; it supports IPv6 for a start)

The stack on what?  Certainly not Linux as it hasn't been around for 30
years.  I would expect any IP stack that was written 30 years ago and
updated as new research provided better insights into the world of
networking to to work very well with VMS.

bill
 

-- 
Bill Gunshannon          |  de-moc-ra-cy (di mok' ra see) n.  Three wolves
billg999 at cs.scranton.edu |  and a sheep voting on what's for dinner.
University of Scranton   |
Scranton, Pennsylvania   |         #include <std.disclaimer.h>