Yes, the bandwidth problems persist while using direct-connections to
hosts, and multiple swapping of new ethernet cables.
I was able to use ttcp and test raw socket connections. Those did not
have any problem getting 80MBps, which is more acceptable.
So the problem is definitely with the Windows applications, which
helps, I guess.
Louis Vitiello Jr. wrote:
> That seems rather odd. Sure the Cisco switch isn't limiting the bandwidth?
>
>
> --
> Louis Vitiello Jr.
> ------------------------------
> MCSE, MCSA, MCP, A+/N+
> ERCP XP Pro / Net Concepts
>
> <(E-Mail Removed)> wrote in message
> news:(E-Mail Removed) ups.com...
> > Running simple tests of Windows Server 2003 Enterprise Edition R2 has
> > yielded an inability to get more than 1Mbps for any single network
> > connection.
> >
> > How may this be tuned to go as fast as possible?
> >
> > I saw these results, using a Gb NIC, on a Cisco Catalyst 6509 with Gb
> > blades and no bandwidth shaping enabled, using modern P4-powered
> > clients with with Gb NICs and 2GB of RAM.
> >
> > The tests were:
> > - large file copies (streaming reads) from shared directories
> > - large message and attachment downloads (again, streaming reads) in
> > a test instance of Exchange, with MAPI and IMAP clients (all MSFT!)
> > - FTP server file downloads (again, streaming reads)
> > - HTTP server file downloads (again, streaming reads)
> >
> > The connection speed was constant and did not vary. 1Mbps per
> > connection the whole-way through. If a client opened multiple
> > connections, another 1Mbps was given to that connection.
> >
> > Surely there is a way for W2k3 to go faster! The hardware certainly is
> > capable of it; this machine and network setup is used to drive test
> > suites for NAS equipment at over 1GBps (NOT 1Gbps--1GBps!) non-stop for
> > weeks.
> >
|
Similar Threads
Linear Mode
