Cisco - LAN Switching

	Internetworking Technologies Instructor: Prabul, CCNA
	Internetworking Technologies Instructor: Prabul, CCNA

2. LAN Switching

Top 10 Graphics:

Ethernet:

a shared-medium technology
all the devices are connected to the same delivery media
media uses a data frame broadcast method
goal is to provide a best effort delivery service
allows all devices to transmit on an equal basis
inherent problem with CSMA/CD technology is collision

CSMA/CD - Carrier sense multiple access with collision detect: Media-access mechanism where devices listen for a carrier before transmitting; if no carrier is sensed for a specific period of time, a device can transmit. If two devices transmit at once, a collision occurs and is detected by all colliding devices. Collisions delay retransmission for a random length of time. CSMA/CD access is used by Ethernet and IEEE 802.3.

IEEE 802.3 - IEEE LAN protocol that specifies an implementation of the physical layer and the MAC sublayer of the data link layer. IEEE 802.3 uses CSMA/CD access over a variety of physical media. Extensions to the IEEE 802.3 standard specify implementations for Fast Ethernet (IEEE 802.3u). Variations of the original IEEE 802.3 specification include 10Base2, 10Base5, 10BaseF, 10BaseT, 10Broad36, 100BaseTX, 100BaseT4, and 100BaseFX.

The performance of a shared media Ethernet/802.3 LAN can be negatively effected by several factors:

The data frame broadcast delivery nature of Ethernet/802.3 LANs
CSMA/CD access methods allow only one station to transmit at a time.
Network congestion due to increased bandwidth demands from multimedia applications such as video and the Internet.
Normal latency (propagation delay) of frames as they travel across the LAN layer 1 media and pass through layer 1, 2 and 3 networking devices.
Extending the distances of the Ethernet/802.3 LANs using Layer 1 repeaters.

Latency:

Delay between the time a device requests access to a network and the time it is granted permission to transmit.
Delay between the time when a device receives a frame and the time that frame is forwarded out the destination port.

Repeater - Layer 1 device that regenerates and propagates electrical signals between two network segments.

Half-Duplex Ethernet Design (Standard Ethernet)

Ethernet physical connector provides several circuits
Most important are receive (RX), transmit (TX), and collision detection

Bandwidth - The difference between the highest and lowest frequencies available for network signals. The term is also used to describe the rated throughput capacity of a given network medium or protocol.

Congestion - Traffic in excess of network capacity. This can be seen in slower response times, longer file transfers and network users becoming less productive due to network delays.

Propagation delay comes from the amount of time it takes a data frame to travel across the physical media and from the time it takes for a frame to pass through a network device such as a repeater. This latency is the total of all the delays caused by the network devices and media as the frame moves from on device to another.

The 'transmission time' is the time it takes the frame or packet to move from the data link layer - the data being placed into a packet or frame, to the physical layer - onto the physical cabling of the network. Each Ethernet bit has a 100ns window for transmission. A byte is equal to eight bits. Therefore, one byte takes a minimum of 800ns to transmit (8 bits at 100ns per equals 800ns). A 64 byte frame takes 51,200ns or 51.2 microseconds to transmit (64 bytes at 800ns equals 51,200ns, 51,200ns/1000 equals 51.2 microseconds).

Attenuation - Loss of communication signal energy. The signal weakens as it travels through the network from the resistance found in the medium.

An Ethernet repeater is a physical layer device on the network that boosts or regenerates the signal on an Ethernet LAN.

repeaters are a Layer 1 device that cleans up and boosts, and passes the signal on
repeaters allow longer end-to-end distances
repeaters increase the collision domain size
repeaters increase the broadcast domain size
repeaters can't filter traffic

This will also compound the issue of broadcasts and collisions and have a negative effect on the overall performance of the shared media LAN.

Hub (multiport repeater)

Generally, a term used to describe a device that serves as the center of a star-topology network.
Hardware or software device that contains multiple independent but connected modules of network and internetwork equipment. Hubs can be active (where they repeat signals sent through them) or passive (where they do not repeat, but merely split, signals sent through them).
In Ethernet and IEEE 802.3, an Ethernet multiport repeater, sometimes referred to as a concentrator.

Improving LAN performance

Segmenting the network (bridges, routers, switches)
Segment:
Section of a network that is bounded by bridges, routers, or switches. Each segment is considered its own collision domain.
- isolate traffic between segments
- achieve more bandwidth per user by creating smaller collision domains
move to Full-duplex

Full duplex:
Capability for simultaneous data transmission between a sending station and a receiving station.
upgrade to Fast Ethernet

Fast Ethernet:
Any of a number of 100-Mbps Ethernet specifications. Fast Ethernet offers a speed increase ten times that of the 10BaseT Ethernet specification, while preserving such qualities as frame format, MAC mechanisms, and MTU. Such similarities allow the use of existing 10BaseT applications and network management tools on Fast Ethernet networks. Based on an extension to the IEEE 802.3 specification.

Collision domain - In Ethernet, the network area within which frames that have collided are propagated. Repeaters and hubs propagate collisions; LAN switches, bridges and routers do not.

Bridge - Device that connects and passes packets between two network segments that use the same communications protocol. Bridges are passive on the network and operate at the 2-Data link layer (layer 2) of the OSI reference model. In general, a bridge will filter, forward, or flood an incoming frame based on the MAC address of that frame.

segmentation provides fewer users per segment
bridges 'store and forward' all frames based on Layer 2 (MAC) addresses
Layer 3 protocol independent
Increase Latency on the network (by 10-30%)
Bridges "learn a network’s" segmentation by building address tables that contain the address of each network device and which segment to use to reach that device.

Store and forward packet switching - Packet-switching technique in which frames are completely processed before being forwarded out the appropriate port. This processing includes calculating the CRC and checking the destination address. In addition, frames must be temporarily stored until network resources (such as an unused link) are available to forward the message.

Interface:

Connection between two systems or devices.
In routing terminology, a network connection.
In telephony, a shared boundary defined by common physical interconnection characteristics, signal characteristics, and meanings of interchanged signals.
The boundary between adjacent layers of the OSI model.

Router - operates at the 3-Network layer and bases all of its forwarding decision between segments on the Layer 3 protocol address. It accomplishes this by examining the destination address on the data packet and looking in its routing table for forwarding instructions. 3-Network layer device that uses one or more metrics to determine the optimal path along which network traffic should be forwarded. Routers forward packets from one network to another based on network layer information. Occasionally called a gateway (although this definition of gateway is becoming increasingly outdated).

more managable, greater functionality, multiple active ports
smaller collision domains
operates at Layer 3 and 4
operate with a higher rate of latency due to acknowledgements
Routers create the highest level of segmentation because of their ability to make exact determinations of where the data packet should be forwarded.

Acknowledgment - Notification sent from one network device to another to acknowledge that some event (for example, receipt of a message) has occurred. Sometimes abbreviated ACK.

Protocols that require an 'acknowledgement' from the receiver to the sender for every packet as it is delivered (known as acknowledgement oriented protocols) have a loss of throughput of 30-40%. Protocols that require minimal acknowledgements (sliding window protocols) suffer a 20-30% loss of throughput.

Switch - Network device that filters, forwards, and floods frames based on the destination address of each frame. The switch operates at the 2-Data link layer of the OSI model. A LAN which uses a switched Ethernet topology creates a network that behaves like it only has two nodes - the sending node and the receiving node.

a switch eliminates the impact of collisions through microsegmentation
low latency (21 msec) and high frame-forwarding rates at each interface port
works with existing 802.3(CSMA/CD) compliant network interface cards and cabling
in a switched Ethernet implementation the available bandwidth can reach closer to 100%.

Because of the switching employed, the MAC address of the destination device is read and the switch begins transmitting the packet before the total packet arrives in the switch. This more than makes up for the inherent latency found in the switch.

Full-duplex Ethernet: Full duplex Ethernet allows the transmission of a packet and the reception of a different packet at the same time. This simultaneous transmission and reception requires two pairs of cables and a switched connection between each node. This connection is considered point-to-point and is nearly collision free. Since both nodes can transmit and receive at the same time there are no negotiations for bandwidth. Full duplex Ethernet can use an existing shared media as long as the media meets minimum Ethernet standards.

significant performance improvement over shared, half-duplex Ethernet
bidirectional traffic (both directions)
uses point-to-point connections - no hubs
collision free

The Full-Duplex Ethernet Switch (FDES) takes advantage of the two pairs of required cables in this configuration by creating a direct connection between the transmit (TX) at one end of the circuit to the receive (RX) end of the other circuit. The same holds true for the opposite circuits on each end. With these two stations connected this way a collision free domain is created. Both stations are not sharing the same wiring pair to transmit and receive.

transmit circuit connects directly to receive circuit
no collisions
significant performance improvement
eliminates contention on Ethernet point-to-point link
uses a single port for each full-duplex connection

Standard Ethernet usually can only use 50-60% of the 10Mbps available bandwidth. This is due to collisions and latency. Full duplex Ethernet offers 100% of the bandwidth in both directions. This produces a potential 20Mbps throughput - 10Mbps TX and 10Mbps RX.

Layer 2 LAN Switch

LAN switches are considered 'multiport bridges' with no 'collision domain' because of 'microsegmentation'. Data is exchanged at high speeds by switching the packet to its destination.

multiport bridge
one broadcast domain
no collision domain
transparent to upper layers
switching and filtering based on Layer 2 MAC addresses
forwards packets based on a forwarding table
enables dedicated access
eliminates collisions and increases capacity
supports multiple conversations at a time

Microsegmentation - Division of a network into smaller segments, usually with the intention of increasing aggregate bandwidth to network devices.

Ethernet switching increases the bandwidth available on a network. It does this by creating dedicated network segments (point-to-point connections) and connecting those segments in a virtual network within the switch. This virtual network circuit exists only when two nodes need to communicate.

Virtual circuit - Logical circuit created to ensure reliable communication between two network devices. A virtual circuit is defined by a VPI/VCI pair, and can be either permanent (a PVC) or switched (an SVC). Virtual circuits are used in Frame Relay and X.25. In ATM, a virtual circuit is called a virtual channel. Sometimes abbreviated VC .

Even though the LAN switch eliminates 'collision domains', all hosts connected to the switch are still in the same 'broadcast domain' . Therefore a broadcast from one node will still be seen by all other nodes connected through the LAN switch.

An Ethernet switch 'learns' addresses by:

reading the 'source address' of each packet transmitted
noting the port where the frame was heard
addind this information to its 'forwarding database'

Addresses are learned 'dynamically' - as new addresses are read they are learned and stored in 'content addressable memory' (CAM). Each time an address is stored it is time stamped and stored for a set period of time. Each time an address is referenced or found in the CAM, it receives a new time stamp. Addresses that are not referenced during set period of time are removed from the list. By removing aged or old addresses the CAM maintains an accurate and functional forwarding database.

Parallelism - Indicates that multiple paths exist between two points in a network. These paths might be of equal or unequal cost. Parallelism is often a network design goal: if one path fails, there is redundancy in the network to ensure that an alternate path to the same point exists.

Symmetrical LAN Switch:

Provides switched connections between like bandwidths (all 10 or all 100 Mbps)
Optimized for even distribution of network traffic across the entire network such as in a 'Peer-to-Peer' environment.

Asymmetric LAN Switch:

Provides switching between unlike bandwidths (10/100 Mbps)
Requires the switch to use memory buffering
Optimized for 'Client-Server' environments

Memory Buffering - An Ethernet switch may use a buffering technique to store-and-forward packets to the correct port or ports. This memory buffer can make use of two methods for forwarding packets

port based memory buffering (packets are stored in queues that are linked to specific incoming ports), or
shared memory buffering (deposits all packets into a common memory buffer that is shared by all the ports on the switch).

'Store and forward' packet switching - Packet-switching technique in which frames are completely processed before being forwarded out the appropriate port. This processing includes calculating the CRC and checking the destination address. In addition, frames must be temporarily stored until network resources (such as an unused link) are available to forward the message.

'Cut-through' packet switching - Packet switching approach that streams data through a switch so that the leading edge of a packet exits the switch at the output port before the packet finishes entering the input port. A device using cut-through packet switching reads, processes, and forwards packets as soon as the destination address is looked up, (first 16 bytes) and the outgoing port determined. Also known as on-the-fly packet switching.

'Fragment Free' packet switching is a variation of 'Cut-through' packet switching with the difference of reading more of the leading edge of the incomming packet (first 64 bytes), enough to be sure the packet is not a bad formatted packet or packet fragment. 'Fragment Free' packet switching does not add appriciably to switching latency.

'Spanning-Tree Protocol' (STP) - Bridge protocol that utilizes the spanning-tree algorithm, enabling a learning bridge to dynamically work around loops in a network topology by creating a spanning tree. Bridges exchange bridge protocol data unit (BPDU) messages with other bridges to detect loops, and then remove the loops by shutting down selected bridge interfaces. Refers to both the IEEE 802.1 Spanning-Tree Protocol standard and the earlier Digital Equipment Corporation Spanning-Tree Protocol upon which it is based. The IEEE version supports bridge domains and allows the bridge to construct a loop-free topology across an extended LAN.

Switches forward broadcast frames
Prevents loops that can cause 'broadcast storms',
(expotentially proliferate frames)
Allows redundant links
Prunes topology to a minimum spanning tree
Resilient to topology changes and device failures

A switch uses STP on all Ethernet and Fast Ethernet-based VLANs. STP detects and breaks loops by placing some connections in a standby mode, which are activated in the event of an active connection failure. A separate instance of STP runs within each configured VLAN, ensuring Ethernet topologies that conform to industry standards throughout the network.

The supported STP states are as follows:

Blocking - No frames forwarded, BPDUs heard
Listening - No frames forwarded, listening for frames
Learning - No frames forwarded, learning addresses
Forwarding - Frames forwarded, learning addresses
Disabled - No frames forwarded, no BPDUs heard

States initially set, later modified by STP
Server ports can be configured to immediately enter STP forwarding mode
Determine the status, cost, and priority of ports and VLANs using the 'show spantree' command.

Unicast - Message sent to a single network destination.

Broadcast - Data packet that will be sent to all nodes on a network. Broadcasts are identified by a broadcast address.

Multicast - Single packets copied by the network and sent to a specific subset of network addresses. These addresses are specified in the destination address field.

VLAN

An Ethernet switch is designed to physically segment a LAN into individual collision domains. However, each segment is still part of one broadcast domain. The total number of segments on a switch equals one broadcast domain. This means that a broadcast from a node on one segment is seen by all nodes on all other segments.

Virtual Local Area Networks (VLANs) are a grouping of network devices or users that is not restricted to a physical switch segment. A VLAN is a logical grouping of devices or users. A VLAN creates a single broadcast domain that is not restricted to a physical segment and is treated like a subnet.

VLAN setup is done in the switch via software. VLANs are not standardized and require the use of proprietary software from the switch vendor.

			`aprabul@yahoo.com` `Copyright © 2003`