RE: [nsp] L3 switching BIGIRON 8000 vs CAT 5500 some testing reports..

From: KF (kf@reign.sk)
Date: Sun Jan 27 2002 - 05:22:29 EST


> You should really compare the BigIron 8k with the
> cat6500 or a cisco7600.

Yes, Sure, but mine current generated traffic wasn't 1gig... but totally 400 Mbit Half duplex.... I did not wanted to compare full
loaded, that's really apples and oranges (have no test bed and for instance Cat 5500 Supervisor III is a 3.6Gbps crossbar switching
fabric ... I wanted to compare L3 switching and QoS under 400Mbit Half duplex..... I think that this is "fair"... anyway

>
> When you measure QoS, you should check out at what
> point does QoS force packets to drop them to the
> ether.. 1st second, 30th second, 60th second? :)
testing still in progress.... anyway my setup for QoS is to compare after QoS applied the L3 switching performance...
>
> What are you using for a packet generator?
Sniffer pro 4.5 and for tcp sessions ftp ...

>
> How much could you push the Gigabit cards under
> Win2k megabyte wise?? Its no where near 1gig...
not in this tests.. I'm not testing GigE... only trunks are GigE.. nevertheless, you have right.. but that's not an question of the
Win2K.. it's about the hardware... try to compare L7 to L7 traffic let's say ftp.... how many Mbit/s you can move from one server to
the another... or let's say 1 server with GigE to 100 clients?... I'm sure that not too much... If you get 300 Mbit/s then your
hardware is really powerfull..

>
> -Joe
>
>
> ---- KF <kf@reign.sk> wrote:
> > Heyla,
> >
> > Would like to share some testing info's wich I
> observed...
> >
> > Big Iron:
> > Version 07.2.09T53
> > Cat5500
> > S-C5500 Software, Version McpSW: 6.3(4) NmpSW: 6.3
> (4)
> > RSM
> > IOS (tm) C5RSM Software (C5RSM-IK8O3SV-M), Version
> 12.2(6a)
> >
> > tested modules:
> > Cat RSM:
> > 12 1 WS-X5302 Hw : 7.5
> > Fw : 20.22
> > Fw1: 3.1(1)
> > Sw : 12.2(6a)
> > Cat Supervisor III
> > 1 0 WS-X5530 Hw : 3.4
> > Fw : 5.1(2)
> > Fw1: 4.4(1)
> > Sw : 6.3(4)
> > Cat 24 port switch
> > 2 24 WS-X5224 Hw : 1.4
> > Fw : 3.1(1)
> > Sw : 6.3(4)
> > 3 gig E SX
> > 5 3 WS-X5403 022127016 Hw : 1.3
> > Fw : 4.5(2)
> > Sw : 6.3(4)
> >
> > Cat 24 port switch
> > 8 24 WS-X5224 Hw : 1.3
> > Fw : 3.1(1)
> > Sw : 6.3(4)
> >
> > Big iron 8000:
> >
> > supervisor with 8 gig E SX
> > S1: BxGMR4 M4 Management Module, SYSIF 2 M4,
> ACTIV
> > switching cards 24 port
> > S2: B24E Copper Switch Module OK
> > S3: B24E Copper Switch Module OK
> >
> > Statement:
> >
> > I agree, that this two devices are not comparable.
> I did comparation in "limits" where both should gave
> as good as they claim
> > performance.
> >
> > Test bed is made of 3 Servers with 100Mbit FDX
> Intel chipset (erghmm compaq neteligent) .. each
> server have around 800 MHZ CPU and
> > High speed SCSI HDDS with STRIPE SET (no parity)
> and 512 MB ram...
> > Servers running Win2K (do not blame me on
> that..they performing very well also with Gigabit
> Eth adapters...have done it years
> > ago...)
> > Netbios and all shitty services disabled...
> >
> > tested with packet generators and ftp (standard
> IIS 5.0 patched and standard ms client)
> >
> > now... I don't want to publish here all data (want
> to hide ;-))) but my results are:
> >
> > - testing L2 switching...
> > Cat 5000 very good
> > Big iron little more performance
> > - testing L3 swithing ....(mls)
> > cat 5000 very good
> > big iron once again better...and better than L2
> switching at all why?!?
> > - testing L3 switching with extended IP access
> list IN/OUT
> > cat 5000 good but CPU RSM running 30 %
> > big iron CPU 5 % and THE SAME throughput...
> 11110.43Kbytes/sec hmmm (output from ftp client..)
> >
> > Now I'm testing QoS... dunno if want to publish
> any nfo's...
> >
> > now...
> >
> > awaiting discussion..from both sides... c'mon...
> let's start hiding..or?
> >
> > Alex
> >
> >
> >
>
>



This archive was generated by hypermail 2b29 : Sun Aug 04 2002 - 04:13:01 EDT