Lec 2. Computer Networks
Lec 2. Computer Networks
Repeaters
A physical layer device the acts on bits not on frames or
packets Can have two or more interfaces When a bit (0,1) arrives, the repeater receives it and regenerates it, the transmits it onto all other interfaces Used in LAN to connect cable segments and extend the maximum cable length extending the geographical LAN range Ethernet 10base5 Max. segment length 500m 4 repeaters (5 segments) are used to extend the cable to 2500m) Ethernet 10Base2- Max. segment length 185m - 4 repeaters (5 segments) are used to extend the cable to 925m Repeaters do not implement any access method If any two nodes on any two connected segments transmit at the same time collision will happen
Hubs
Acts on the physical layer Operate on bits rather than frames Also called multiport repeater Used to connect stations adapters in a physical star topology but logically bus Connection to the hub consists of two pairs of twisted pair wire one for transmission and the other for receiving. Hub receives a bit from an adapter and sends it to all the other adapters without implementing any access method. does not do filtering (forward a frame into a specific destination or drop it) just it copy the received frame onto all other links The entire hub forms a single collision domain, and a single Broadcast domain
Collision domain: is that part of the network (set of NICs) when two or
more nodes transmit at the same time collision will happen. Broadcast domain: is that part of the network (set of NIC) where each NIC can 'see' other NICs' traffic broadcast messages.
Multiple Hubs can be used to extend the network length For 10BaseT and 100BaseT the maximum length of the connection
between an adapter and the hub is 100 meters the maximum length between any two nodes is 200 m = maximum network length
Figure 16.4
Hubs
Disadvantages
Individual segment collision domains become one large collision
domain (reduce the performance) Cant interconnect different Ethernet technologies(like 10BaseT & 100BaseT) because no buffering at the hub
Here we have a single collision domain and a single broadcast domain
information about the network and report them to a monitoring host connected to the hub so some statistics about the network (bandwidth usages, collision rates, average frame sizes) can be generated. If an adapter is not working the hub can disconnect it internally and the network will not be affected.
Bridges/switches
Acts on the data link layer (MAC address level) Used to divide (segment) the LAN into smaller LANs segments, or to connect LANs that use identical physical and data link layers protocol (see figure in next slide) Each LAN segment is a separate collision domain Bridge does not send the received frame to all other interfaces like hubs and repeaters, but it performs filtering which means:
Whether a frame should be forwarded to another interface that leads to the destination or dropped
This is done by a bridge table (forwarding table) that contains entries for the nodes on the LAN The bridge table is initially empty and filled automatically by learning from frames movements in the network An entry in the bridge table consists of : Node LAN (MAC) Address, Bridge Interface to which the node is connected to, the record creation time
A bridge runs CSMA/CD before sending a frame onto the link not like the hub or repeater Bridge frame handling is done in software
Bridges
the frame was received if it is different from the one in the table also it updates the record time
Then, the switch compares the destination address of the frame with
The bridge compares the interface number on which the frame was received and the interface number in the table, if they are different the bridge forwards the frame through the interface number stored in the table. Otherwise, if they are the same the switches discards (drops) the frame.
If no match is found, the switch floods the frame on all interfaces except
In heavy load, each station has an average effective theoretical bandwidth = 10/12
Switch
Example: Three LANs connected through a bridge Note: here we have three collision domains and a single broadcast domain
Figure 16.8
Switch
When using switches, the network should not contain any loop (there should be exactly one path from any LAN to any other LAN Loops can cause number of frames in the LAN to increase indefinitely
Switches
N-Port bridge where N is equal to number of stations Usually used to connect individual computers not LANs like bridge Allows more than one device connected to the switch directly to
transmit simultaneously Can operates in Full-duplex mode (can send and receive frames at the same time over the same interface) Performs MAC address recognition and frame forwarding in hardware (bridge in software) Two types : Store-and-forward: switch receives the whole a frame on the input line, buffers it briefly , performs error checking, then routes it to the appropriate output line (similar to bridge). Buffering will cause some delay. Cut-through: based on the fact that the destination address appears at the beginning of the MAC frame, so once the address is recognized the frame is directly sent to the appropriate output line if the output buffer is empty (no need to buffer it). no buffering delay NO ERROR CHECKING
Full-Duplex operation
Routers
Operates at network layer = deals with packets not frames Connect LANs and WANs with similar or different protocols
together Switches and bridges isolate collision domains but forward broadcast messages to all LANs connected to them. Routers isolate both collision domains and broadcast domains Acts like normal stations on a network, but have more than one network address (an address to each connected network) Deals with global address ( network layer address (IP)) not local address (MAC address) Routers Communicate with each other and exchange routing information Determine best route using routing algorithm by special software installed on them Forward traffic if information on destination is available otherwise discard it (not like a switch or bridge)
Routers
switch
switch
switch
Learning Objectives
List common cable types used in networking
Describe how UTP cables are made Explain how UTP cables are used in Ethernet networks
Unshielded
twisted pair
Fiber optic
UTP characteristics
Unshielded Twisted (why?) pairs of insulated conductors
Covered by
insulating sheath
UTP categories
Category 1 Category 2 Category 3 Category 4 Category 5 Category 5e Voice only (Telephone) Data to 4 Mbps (Localtalk) Data to 10Mbps (Ethernet) Data to 20Mbps (Token ring) Data to 100Mbps (Fast Ethernet) Data to 1000Mbps (Gigabit Ethernet)
Category 6
Cat5e cable
1000Mbps data capacity
For runs of up to 90 meters Solid core cable ideal for structural installations (PVC
or Plenum) Stranded cable ideal for patch cables Terminated with RJ-45 connectors
RJ45 connector
Cable stripper
Scissors Crimping tool
Step 6 - Check
Do all wires extend to end? Is sheath well inside connector?
Step 7 - Crimp
Squeeze firmly to crimp connecter