From: CCIE 2004 (ccie2004@excite.com)
Date: Mon May 17 2004 - 19:25:25 GMT-3
Hi Howard,
Thank you for your response. However I had another question. Per your response
"Bridges and layer 2 switches make decisions on _destination_ MAC
addresses. If VLANs are used, layer 2 switches may make their first
decisions, on a trunk port, on the VLAN ID field."
Could you please confirm as from whatever I have seen and read it always says that Bridges and Switches build their tables based on the source mac address. I am not sure what you were referring to and I might have interpreted it the wrong way.
Also, I was trying to find out how others would define Layer 3 switching and differentiate it with traditional Layer 2 switching so that I could gain a better understanding of what it typically means or is it just marketing hype and a layer 3 switch is basically a layer 2 switch with say either a RSM like the Catalyst 5000s or a MSFC like for the catalyst 6000. They are traditionally layer 2 switches with just routing functionality.
Thanks for your assistance.
--- On Sat 05/15, Howard C. Berkowitz < hcb@gettcomm.com > wrote:
From: Howard C. Berkowitz [mailto: hcb@gettcomm.com]
To: ccielab@groupstudy.com
Date: Sat, 15 May 2004 10:27:39 -0400
Subject: Re: Difference between Layer 2 and Layer 3 switches
At 8:57 AM -0400 5/15/04, CCIE 2004 wrote:
> Thanks for your reply Chris. Does anyone have any other thoughts. Thanks ---
The first thought would be whether you are talking about bridges
versus routers, which are well-defined technical terms, or layer 2
switches versus layer 3 switches, which are largely marketing terms
dating back to several vendors' slogans "switch when you can, route
when you must."
There's actually more of a distinction that really can be made
between a traditional bridge and even a basic layer 2 switch. The
switch, in this case, is intended to microsegment -- use one physical
port to connect to each end device, removing, as long as the
connectivity is full duplex, collisions. Switches also are more
likely to support VLAN trunking than traditional bridges, which are
interconnected simply as part of a homogeneous spanning tree.
>On Tue 05/11, Chris Larson < clarson52@comcast.net >
>wrote:From: Chris Larson [mailto: clarson52@comcast.net]To:
>ccie2004@excite.comDate: Tue, 11 May 2004 06:59:47 -0400Subject: Re:
>Difference between Layer 2 and Layer 3 switches There is a lot too
>it. But in a pure layer 2 switch, to route between vlansyou will
>need an external router. In a layer 3 switch, routing can be
>donewithin the switch. Also, a layer 3 switch is probably going to
>give you QoSabilities like traffic marking and cvlassificagtion
>based on layer 3 andlayer 4 info in the packet.
Actually, it's perfectly common and useful to have layer 4, but not
layer 3, capabilities in something one might consider a layer 2
switch. An increasingly common carrier application, when metro
Ethernet services are deployed, is to put such a switch on the
customer premises. In this application, the switch will have a port
for each VPN and for Internet connectivity, as well as a port for
voice (and possibly video). All of these ports are mapped into
802.1q VLANs for transmission to the carrier POP, where they are then
mapped to routed VPNs. At the customer premises switch, 802.1p QoS
bits are set on the voice/video VLANs, which variously are mapped to
traffic-engineered VPNs at the POP L3 function, and/or have 802.1p
mapped to DSCP in IP packets.
>----- Original Message ----- From: "CCIE 2004" To: Sent: Tuesday,
>May 11, 2004 12:54 AMSubject: OT: Difference between Layer 2 and
>Layer 3 switches> Hi, I just wanted to get a brief overview of
>the general concept of layer2 and layer 3 switches. I did read a
>thread earlier on groupstudy and hadseen a recent article but cannot
>remember where I saw !
> the article. I thinklayer 3 switches are mainly routers that are
>capable of caching and doing alookup on network layer addresses (ip,
>ipx etc) in Asic.
This is a really old definition that Cisco persists in continuing.
ASIC is not a very precisely defined characteristic. I can build a
router with intelligent forwarding cards, each with a general purpose
RISC processor, that will hammer a single-forwarding-ASIC router into
the ground.
An ASIC may be a pure hardware device like a fixed gate array, but is
probably more likely to be at least somewhat flexible, such as a
field-programmable (e.g., through EEPROM) gate array (FPGA), a
microcode sequencer (again with code either burned into ROM or in an
EEPROM), or even a general-purpose RISC processor with the
instructions prefetched, decoded, and in ROM/EEPROM.
>Layer 2 switchesare mainly devices that build their table on source
>mac address.
Bridges and layer 2 switches make decisions on _destination_ MAC
addresses. If VLANs are used, layer 2 switches may make their first
decisions, on a trunk port, on the VLAN ID field.
This archive was generated by hypermail 2.1.4 : Wed Jun 02 2004 - 11:12:13 GMT-3