cisco | Page 4

switch runs at Layer 2 and cannot learn t he MAC address FFFFFFFFFFFF.
A collection of interconnected switches forms a bro adcast domain.
Only Layer 3 devices or a VLAN form separate broadc ast domains.
Design Considerations – Ethernet/802.3
Network Latency:
Latency is the time a frame or a packet takes to tr avel from the source to the final destination.
Network Congestion:
The primary reason for segmenting a LAN into smalle r parts is to isolate traffic and to achieve better use of
bandwidth per user.
Without segmentation, a LAN quickly becomes clogged with traffic and collisions.
Most common causes:
Increasingly powerful computer and network technolo gies.
Increasing volume of network traffic.
High-bandwidth applications.
LAN Segmentation:
LANs are segmented into a number of smaller collision and broadcast domains using routers and switches.

LAN Design Considerations
There are two primary considerations when designing a LAN:
Controlling network latency
Removing bottlenecks
Controlling Network Latency:

Consider the latency caused by each device on the n etwork.(switch)
Switches at Layer 2 can introduce latency on a network when oversubscribed on a busy network.
If a core level switch has to support 48 ports, eac h one capable of running at 1000 Mb/s full duplex, the switch
should support around 96 Gb/s internal throughput i f it is to maintain full wire speed across all ports simultaneously.
Consider the latency caused by each device on the n etwork.(router)
The use of higher layer devices can also increase l atency on a network.
When a Layer 3 device, such as a router, needs to e xamine the Layer 3 addressing information contained within the
frame, it must read further into the frame than a Layer 2 device, which creates a longer processing time.
Removing Network Bottlenecks:
Each workstation and the server are connected at 1 000Mbps.

Switch Concepts and Configuration
Switch Forwarding Methods
Methods switches use to forward Ethernet frames.
Store-and-forward.
Cut-through:
Fast-forward switching.
Fragment-free switching.
-Store-and forward:
Receives the entire frame.
Computes the CRC and checks the frame length.
If valid, checks the switch table for the destinati on address and forwards the frame.
If invalid, the frame is dropped.
Store-and forward is the only method used on curren t Cisco Catalyst switches.
Needed for QoS on converged networks.
-Cut-through:
Forwards a frame before it is entirely received.
At a minimum, it must read the destination and sour ce MAC addresses.
Faster than store-and-forward.
No error checking.
Any corrupt frames are still forwarded and consume network bandwidth.
-Cut-through – Fast-forward:
Typical method of cut-through.
Forwards a frame immediately after it reads and fin ds the destination address.
-Cut-through – Fragment-free:
Stores the first 64 bytes of the frame before forwa rding.
The first 64 bytes of the frame is where most netwo rk errors and collisions occur.

Checks for a collision before forwarding the frame.
Some switches are configured to use cut-through on each port until a user defined error threshold is reached. At
that time, they change to store-and forward.
Symmetric and Asymmetric Switching
Symmetric:
All ports are of the same bandwidth.
Optimized for a reasonably distributed traffic load .
For example, a peer-to-peer network.
Asymmetric:
Provides switched connections between ports
of unlike bandwidth.
For example, more bandwidth can be assigned to a se rver to prevent bottlenecks.



Memory Buffering
A switch analyzes some or all of a packet before it forwards it to the destination host based on the forwarding
method.
It stores the packet for the brief time in a memory buffer.
Built into the hardware
Two types:
Port based.
Shared.
Port Based:
Frames are stored in queues that are linked to spec ific incoming and outgoing ports.
A frame is transmitted to the outgoing port only when all the frames ahead of it in the queue have been
successfully transmitted.
It is possible for a single frame to delay the tran smission of all the frames in memory because of a b usy destination
port.
Shared:
Deposits all frames into a common memory buffer tha t all the ports on the switch share.
The amount of buffer memory required by