What is up with the pricing for gigabit ports on Cisco 3000 series
switches? On the 3500 line prices per-port are 3x to 5x that for
equivalent ports on H-P enterprise-class switches (e.g. 5308XL-48G).
Even the 3700 series is priced higher than most similar competitors.

Gigabit technology has been on the general market for 5-6 years now and
even Layer 3 Gigabit switching is becoming a commodity. Is there
something magical about Cisco gigabit ports I am missing in the specs
that makes them worth so much more?

sPh

Users who have bought HP ProCurve really like them:

http://www.bradreese.com/hp-procurve...ory-search.htm

Most engineers are trained on Cisco.

This provides Cisco with pricing leverage.

However, responding to customer pricing concerns, Cisco has rolled-out
their "Factory Refurbished Program" ( warranty same as Cisco New as
well as SMARTnet Eligible ).

Daily Update SMARTnet Eligible Cisco Refurbed Inventory Availability:

http://www.bradreese.com/cisco-inventory-search.htm

You may wish to investigate the Cisco SMARTnet Pricing Calculator too:

http://www.bradreese.com/cisco-smartnet.htm

Sincerely,

Brad Reese
BradReese.Com Cisco Repair Worldwide
United Kingdom: 44-20-70784294
U.S. Toll Free: 877-549-2680
International: 828-277-7272
Fax: 775-254-3558
Website: http://www.bradreese.com/cisco-big-iron-repair.htm

In article <1106240924.635437.152960@z14g2000cwz.googlegroups .com>,
sphealey <sphealey@worldnet.att.net> wrote:
:What is up with the pricing for gigabit ports on Cisco 3000 series
:switches? On the 3500 line prices per-port are 3x to 5x that for

Just as a small matter of semantics: although Cisco uses an initial
model number of '3' to designate many of its fixed-configuration
LAN switches ('2' for some of them), they are not a "series".
One has to get down as far as "3550" to designate the series you
are probably referring to, given your reference to "3700".
Without that reference, you might have been talking about the
"3500 XL series", which is quite different from the 3550 series.

:equivalent ports on H-P enterprise-class switches (e.g. 5308XL-48G).
:Even the 3700 series is priced higher than most similar competitors.

Including close to 4 times the price per port of Nortel's 5500's.

:Gigabit technology has been on the general market for 5-6 years now and
:even Layer 3 Gigabit switching is becoming a commodity. Is there
:something magical about Cisco gigabit ports I am missing in the specs
:that makes them worth so much more?

We looked at the Nortel 5510, considering it's much lower price,
but we still ended up with a Cisco 3750. Nortel's 5510 will -some day-
be a competitor for what we needed, but it is not there yet.
The 5510 has only had layer 3 switching since May 2004, and as of
November still only permitted static routes.

According to the data sheets, the HP Procurve 5300 series supports
static routes, RIP, RIPv2, and OSPF... but not BGP. According to
the reference documentation, the 5300 does not support policy-based
routing or private vlans.

QoS priotization on the 5300 can be by IP address (which will apply
whether the IP is the source or destination, you can't be specific), or
by TCP/UDP port number [destination I guess, the manual isn't
specific], but there is no way to specify an 'and' of those two to get
the equivilent of a Cisco "extended" ACL entry. Oh yes, the port
numbers and IP numbers must each be specified individually -- you
cannot specify port ranges or lists, and you cannot even specify a mask
for the IP address. If you want a different interface or vlan to have
a different priortization for a given TCP port (say), then you can't
do that -- there is a rigid hierarchy in which the classification
criteria are applied, and TCP/UDP port number is always #1,
vlan # is always #5, interface # is always #6.

The only rate limiting on the Procurve is per-port, and the only limit
available is to specify the percentage of the carrying capacity of the
link that the inbound traffic is limited to.

"For example, if a 100 Mbps port negotiates a link at 100 Mbps and
is rate-limit configured at 50%, then the inbound traffic flow through
that port is limited to no more than 50 Mbps. Similarily, if the
same port negotiates a 10 Mbps link, then it allows no more than
5 Mbps of inbound traffic."

(Page 13-10 of the Feb 2004 reference manual.)


In other words, the "magic" in the 3550 and 3750 series is that they
are much more flexible than the HP Procurve 5300XL or Nortel Baystack 5500
series. Some day, the Procurve or Baystack lines may catch up to
where Cisco 3550 or 3750 have now, but if you need that flexibility
*now*, you can't wait for HP or Nortel.
--
Cannot open .signature: Permission denied

OK, sorry about that. I was trying to keep things a bit generic. I am
looking at the 3550 and 3750 models in general and the 3550-24-FX,
3550-24-SMI, and 3550-12T specifically.

That by the way is one of my concerns about Cisco: the tremendous
number of overlapping and in some cases contradictory models and
product lines not only makes it hard to figure out what might meet my
needs, but raises questions as to how soon any specific model might
evaporate.

The other concern is having to build a fairly complex stack with more
boxes to manage and more power supplies to fail rather than having a
single chassis to manage.
Have you looked at the 8.0 software for the H-P 5300 yet?

sPh

In article <1106253389.758808.111590@z14g2000cwz.googlegroups .com>,
sphealey <sphealey@worldnet.att.net> wrote:
:Have you looked at the 8.0 software for the H-P 5300 yet?

Yup, all of my notes about the Procurve were based on reading the
latest available manual and release notes on HP's site. For example,
rate limiting for other than ICMP did not get added until after the
first 8.0 release.
--
Preposterous!! Where would all the calculators go?!

sphealey wrote:

Yes. They have the sticker with "Cisco" on them!

--

hsb


"Somehow I imagined this experience would be more rewarding" Calvin
**************************ROT13 MY ADDRESS*************************
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account. Please post a followup instead.
************************************************** ******************

sphealey wrote:
Or rolling out "Enterprise class" switches w/o full snmp support for
the values that enterprises would care about!

They are trying to stick by the "5 year" end of sale policy for the
newly rolled out boxes.

Look carefully at management issues. The commands are slightly
different and the stack management may be an issue for your operations
team.

--

hsb


"Somehow I imagined this experience would be more rewarding" Calvin
**************************ROT13 MY ADDRESS*************************
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account. Please post a followup instead.
************************************************** ******************

HSB,
Could you expand on that a bit? Were you referring to Cisco or H-P?
Thanks.

sPh

sphealey wrote:
Sure. The stackable 3700 series with PoE switches were the ones I was
referring to. If you look at what's available via snmp, it's quite
lacking. In fact, we punted on this switch due to lack of what I call
"enterprise level" snmp support.

--

hsb


"Somehow I imagined this experience would be more rewarding" Calvin
**************************ROT13 MY ADDRESS*************************
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account. Please post a followup instead.
************************************************** ******************

In article <jOfId.61833$kq2.48960@twister.nyc.rr.com>,
Hansang Bae <uonr@alp.ee.pbz> wrote:
:Sure. The stackable 3700 series with PoE switches were the ones I was
:referring to. If you look at what's available via snmp, it's quite
:lacking. In fact, we punted on this switch due to lack of what I call
:"enterprise level" snmp support.

As someone at the lower boundary of "enterprise level" I would
be interested in a "for-instance"; I find that a lot of my learning
comes from hearing about people doing things that I never even thought
about doing.
--
How does Usenet function without a fixed point?

Walter Roberson wrote:
Here's a synopsis of some of the findings by my colleagues.:

1) When powering off the existing master switch, PINGs will drop 3-4
packets but SNMP agent will report down for about 30 seconds. This will
potentially generate false alarms, as the SNMP hosts will keep polling
the agent.

2) After removing a stack member, all the relevant member Ethernet
ports will stay with the master and still can be configured with all
commands, even after reboot. This creates an issue, as SNMP host
reports the removed switch ports with status of "admin up" (hardware
present) & "operation down" (not connected). If SNMP is configured to
keep polling the ports of the removed stack member, we'll continue to
get false alarms (unless we go in and run the "shutdown" commands on
the removed interfaces).

More in-depth study reveals that "IF Table" MIB will report all these
ports in "down down" state(but not "gone"). To the contrary, the STACK
MIBs reported the group of ports were removed. So conflicting info was
generated from different MIBs. If we telnet into the Master switch,
"show interface" command will give "hardware not present" under the
removed ports. But from SNMP alone, we can't really tell whether the
ports are gone or just not connected to PCs.

There were other issues that I didn't dig up to post.

--

hsb


"Somehow I imagined this experience would be more rewarding" Calvin
**************************ROT13 MY ADDRESS*************************
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account. Please post a followup instead.
************************************************** ******************