S3C2 – LAN Switching

Download Report

Transcript S3C2 – LAN Switching

S3C2 – LAN Switching
Addressing LAN Problems
Congestion is Caused By
•
•
•
•
Multitasking,
Faster operating systems,
More Web-based applications
Client-Server
– client/server applications allow administrators to
centralize information, thus making it easy to maintain
and protect.
• Point-to-point (host to host) connection is collision
free
LANS Impacted By:
• The data frame broadcast delivery nature of
Ethernet/802.3 LANs
• Carrier sense multiple access collision detect
(CSMA/CD) access methods allowing only one
station to transmit at a time
• Multimedia applications with higher bandwidth
demand such as video and the Internet, coupled
with the broadcast nature of Ethernet, can create
network congestion.
Impact Issues Continued
• Normal latency as the frames travel across
the Layer 1 medium and through Layer 1, 2,
and 3 networking devices, and the latency
added by the extension of Ethernet/802.3
LANs by adding repeaters
• Extending the distances of the
Ethernet/802.3 LANs by using Layer 1
repeaters
Latency
• Latency is caused by:
– the time it takes the source NIC to place voltage
pulses on the wire and the time it takes the
receiving NIC to interpret these pulses. This is
sometimes called NIC delay (it is typically
around 1 microsecond for10BASE-T NICs).
– A byte takes a minimum of 800 ns to transmit
Latency Continued
– Second, there is the actual propagation delay as the
signal takes time -- albeit a very short time -- to actually
travel down the cable (it is typically about .556
microseconds per 100 m for Cat 5 UTP). The longer the
cable, the more propagation delay; the slower the
nominal velocity of propagation (NVP) of the cable, the
more the propagation delay.
– Third, latency is added according to which networking
devices -- whether they be Layer 1, 2, or 3 (and how
they are configured) -- are added in the path between
the two communicating computers. The actual
transmission time (the duration of the host actually
sending bits) must also be included in understanding
timing on networks.
Half Duplex or Duplex
• Ethernet LANs are half-duplex technology
• Full-duplex Ethernet allows the transmission of a
packet and the reception of a different packet at
the same time – requires full duplex NIC card
– This simultaneous transmission and reception requires
the use of two pairs of wires in the cable and a switched
connection between each node. This connection is
considered point-to-point and is collision free.
– Because both nodes can transmit and receive at the
same time, there are no negotiations for bandwidth.
Full-duplex Ethernet can use an existing shared
medium as long as the medium meets minimum
Ethernet standards. Requires 2 pairs of wires and
switched connection
Benefit of Full Duplex
• Ethernet usually can only use 50%-60% of the 10Mbps available bandwidth because of collisions
and latency.
• Full-duplex Ethernet offers 100% of the
bandwidth in both directions.
• This produces a potential 20-Mbps throughput10-Mbps TX and 10-Mbps RX.
• Remember – Transmit connects to Receive
– Think SIMPLE LAN
Why Segment
• Isolates certain traffic
• Creates smaller collision domains
• Data are passed between segments using
bridges, switches, or routers
• IMPORTANT
– Decreasing size of collision domains increases
the number of collision domains
Bridges
• Create tables to match segments and MAC
addresses
• Layer 3 protocol independent
• Stores and then forwards based on MAC
address
• Increases latency on network (10-30%)
• Creates smaller collision domains but
increases number of collision domains
Routers
• Operate at Layer 3 using IP addresses
• More manageable, greater functionality, multiple
paths
• Smaller collision domains
• Introduce latency
– Protocols that provide acknowledgements introduce
30-40% delays
– Protocols that provide minimal acknowledgements have
20%-30% loss in throughput
– So routers introduce 20-40% delay depending on
protocol
Switches/Switched Ethernet
• Low latency and high frame-forwarding rates
• Eliminates impact of collisions through microsegmentation
• Works with existing 802.3 standards
• Create dedicated network segments (point to
point) with full bandwidth – virtual circuits
• Create collision free domains
• Cost more than bridges or routers
Bridges vs Switches
• Both bridges and switches connect LAN
segments, use a table of MAC addresses to
determine the segment on which a datagram needs
to be transmitted, and reduce traffic.
• Switches are more functional in today’s networks
than bridges because they operate at much higher
speeds than bridges and can support new
functionality, such as virtual LANs (VLANs).
• Bridges typically switch using hardware; switches
typically switch using software.
Layer 2 Switching
• With Layer 2 switching, frames are switched
based on MAC address information.
• If the Layer 2 switch does not know where to send
the frame, it broadcasts the frame out all its ports
to the network to learn the correct destination
• the switch learns the location of the new address
and adds the information to the switching table
– MAC Address and ports
Virtual LANs
• dedicated paths between sending and
receiving hosts within the switch are
temporary. The switch’s power comes from
the fact that it can rapidly make and break
these 1 to 1 connections through its
various ports, depending upon the data in
its switching table.
How A Switch Learns Addresses
• Examines source address
• Sends out all ports except incoming port when
address is unknown, multicast, or broadcast
• Forwards when the destination is at a different
interface
• Filters when the destination is on the same
interface
• Date stamps each address– discards after a certain
time period
• Addresses stored in CAM – Content Addressable
Memory
Benefits of Switching
•
•
•
•
•
Number of collisions reduced
Simultaneous multiple communications
High speed uplinks
Improved network response
Increased user productivity
Symmetric/Asymmetric
Switching
• Symmetric switching provides switching
between like bandwidths
– Multiple simultaneous conversations increase
throughput
• Asymmetric provides switching between
unlike bandwidths
– Requires the switch to use memory buffering
Spanning Tree Protocol
• Switches forward broadcast frames
• Prevents loops
– Loops can cause broadcast storms and exponentially
proliferate fragments
•
•
•
•
•
•
Allows redundant links
Prunes topology to a minimal spanning tree
Resilient to topology changes and device failures
Spanning Tree Frames are called bridge
protocol data units (BPDUs)
Spanning Tree enabled by default on catalyst
switch
Spanning Tree States
• States are initially set and then modified by STP
–
–
–
–
–
Blocking
Listening
Learning
Forwarding
Disabled
• Server ports can be configured to immediately
enter SPT forwarding mode
• You can determine the status, cost, and priority of
ports and VLANs by using the show spantree
command
Switching Modes
• Store and Forward
– Entire frame received before forwarding takes place –
causes more latency but error detection is high
• Cut Though (Read first 6 bytes)
– Switch reads destination address before receiving entire
frame and it is forwarded – decreases latency but higher
error rate
• Fast forward immediately forwards
• Fragment Free filters out collision fragments
• Fragment Free (Read first 64 bytes)
– Ensures frame is not a runt and probably not an error
Buffering
• Two methods
– Port-based
• Packets stored in queues that are linked to incoming ports –
packets forwarded when queue is clear
– Shared Memory buffering
• Deposits all packets into common memory buffer shared by all
ports
– Dynamic location assigns port areas
– Switch maintains a map of ports and clears when packet
is switched
Virtual LAN
• Logical network independent of their members’
physical locations
• Administratively defined broadcast domain
• Users reassigned to different VLAN using
software
• Can be grouped by function, department,
application
• Creates a single broadcast domain that is not
restricted to physical segment
• INSTANT LEVEL OF SECURITY