Hubs/Repeaters
Also known as concentrators or repeaters, Hubs crossover and repeat incoming signals to all outgoing lines. They do not redirect or filter any of the data that they receive. They simply regenerate incoming packet to all their ports. Bridges, switches and routers are all types of hubs that perform specialized functions within the network.
TypesClass I – data comes in translated to digital – retranslated to line signal before being sent out. This allows for mixing of 100baseTX and 100baseFX which use different signaling techniques. Because of the timing delay involved in the translation from line to digital it is not possible to connect two class I hubs together. Nor is it possible to connect a class one hub to a class two hub.
Class II – There is no signal translation therefore network types cannot be mixed. Class II Fast Ethernet hubs may be directly connected together. There is a limit to how many hubs may be connected together. The Ethernet standard limits the time that a signal may be propagated throughout a network. Since hubs slow the transmission of a packet, only a limited amount may be used within one network.
Repeaters
Because of electric attenuation in copper wires, there is a limit to the distance allowed between two computers before signals passing between them become unusable. Repeaters regenerate the signals and increase the maximum allowable distances between two computers. Some hubs perform this regenerative function on the electric signals, others don’t.
Crossover
In order to communicate two computers must have the signals crossed so the send wire from one computer enters the receive wire of the other. Hubs cross over incoming electric signals internally. Two hubs can only communicate if their signals are uncrossed or straight through. The same is true for a direct connection to a server or end station.
Wiring
Straight through cables are used to connect a computer to a hub. Crossover cables are used to connect a hub to a hub. A straight through cable may be used to connect a hub to a hub or to a server if the hub has a special uplink port. The MDI/MDI-X push-button switch must be set to the uplink position.
Hub cabling Rules
The total length between two computers can't exceed 205m
A + B + C < 205
Each cable can't exceed 100m.
A,B,C < 100
Cable connecting two hubs can't exceed 5m
B < 5
TypesClass I – data comes in translated to digital – retranslated to line signal before being sent out. This allows for mixing of 100baseTX and 100baseFX which use different signaling techniques. Because of the timing delay involved in the translation from line to digital it is not possible to connect two class I hubs together. Nor is it possible to connect a class one hub to a class two hub.
Class II – There is no signal translation therefore network types cannot be mixed. Class II Fast Ethernet hubs may be directly connected together. There is a limit to how many hubs may be connected together. The Ethernet standard limits the time that a signal may be propagated throughout a network. Since hubs slow the transmission of a packet, only a limited amount may be used within one network.
Repeaters
Because of electric attenuation in copper wires, there is a limit to the distance allowed between two computers before signals passing between them become unusable. Repeaters regenerate the signals and increase the maximum allowable distances between two computers. Some hubs perform this regenerative function on the electric signals, others don’t.
Crossover
In order to communicate two computers must have the signals crossed so the send wire from one computer enters the receive wire of the other. Hubs cross over incoming electric signals internally. Two hubs can only communicate if their signals are uncrossed or straight through. The same is true for a direct connection to a server or end station.
Wiring
Straight through cables are used to connect a computer to a hub. Crossover cables are used to connect a hub to a hub. A straight through cable may be used to connect a hub to a hub or to a server if the hub has a special uplink port. The MDI/MDI-X push-button switch must be set to the uplink position.
Hub cabling Rules
The total length between two computers can't exceed 205mA + B + C < 205
Each cable can't exceed 100m.
A,B,C < 100
Cable connecting two hubs can't exceed 5m
B < 5
Bridges
One of the physical properties of an Ethernet is that all computers share the same electric wire. Therefore computers compete to have access of the wire for communicating. If two computers attempt to transmit data at the same time a collision event occurs and the computers keep vying for control of the Ethernet until one wins out. It becomes obvious that collisions increase with the quantity of computers on a LAN to the point where the entire network can be drastically bogged down. Bridges attempt to solve this problem be dividing the network in separate collision domains.
Mixed Media / Network Types
Bridges have the ability to link two networks of different media types. They are able to translate data into a number of different frame formats making communications among dissimilar networks possible. Some routers have translation bridging software imbedded, therefore eliminating the need for bridges.
Another benefit of bridges is that they can connect network segments of different media and network types, such as 10baseT and 100baseT. Bridges disregard the network protocol (TCP/IP, IPX, etc.) being used since they act on the data-link level (OSI layer 2). They act as simple gates, basing their decisions to let data pass or not on the hardware address of the individual Ethernet cards. The bridge looks at the destination and source addresses of individual packets, if these are within the same segment or collision domain the data is dropped. If the data has a destination address on a different network segment (the other side of the bridge), the data is let it through.
Address mapping
Bridges map MAC addresses by listening to network traffic and require no programming. This autonomous behavior increases the probability of errors if too many bridges are used on a network. A situation in which a bridge forwards all data because of confusion about the location of nodes can be avoided by adhering to the Spanning Tree Algorithm.
Mixed Media / Network Types
Bridges have the ability to link two networks of different media types. They are able to translate data into a number of different frame formats making communications among dissimilar networks possible. Some routers have translation bridging software imbedded, therefore eliminating the need for bridges.
Another benefit of bridges is that they can connect network segments of different media and network types, such as 10baseT and 100baseT. Bridges disregard the network protocol (TCP/IP, IPX, etc.) being used since they act on the data-link level (OSI layer 2). They act as simple gates, basing their decisions to let data pass or not on the hardware address of the individual Ethernet cards. The bridge looks at the destination and source addresses of individual packets, if these are within the same segment or collision domain the data is dropped. If the data has a destination address on a different network segment (the other side of the bridge), the data is let it through.
Address mapping
Bridges map MAC addresses by listening to network traffic and require no programming. This autonomous behavior increases the probability of errors if too many bridges are used on a network. A situation in which a bridge forwards all data because of confusion about the location of nodes can be avoided by adhering to the Spanning Tree Algorithm.
Routers
Routing involves two activities: determining an optimal path and transporting packets. Routers work on level 3 of the OSI model. They use network addresses to map their environment.
Address Translation
A computer’s true network address is its MAC address or hardware address. This series or numbers is hardwired into the computer’s network interface card and is a unique number that distinguishes it from all the other network cards in the world. Unfortunately addressing data packets using the MAC address imposes serious restrictions on network size because there is no facility to route packets using hardware addressing. IP addressing overcomes this restriction by its ability to combine computers and networks into logical groups. Because of the hierarchical structure of IP addressing each router is able to either, determine the destination computers MAC address or find a router that could help bring the packet one step closer to one that can.
There needs to be some mechanism in place to translate IP addresses to hardware addresses for an IP datagram to be transmitted and received. ARP (Address Resolution Protocol) stipulates that when a TCP/IP computer first comes on the network it is to advertise its presence by transmitting its hardware and IP addresses. The other systems then use this information to update their translation address tables (ARP Cache). If a computer is asked to send data to a destination that does not have an entry in its ARP Cache, the originating computer sends out a broadcast message (ARP Request) to every computer on the network asking the destination computer to announce itself. If the destination computer announces its presence, the ARP cache is updated and the datagram is sent on its way. If the destination computer resides on a different network or subnet, the datagram is forwarded to a router.
The data packet will contain the target computers “network” or IP address and since it doesn’t know the target computers MAC address, the first routers MAC address will be substituted instead.
If the first router(router 1) cannot resolve the network address and determines that the packet must be forwarded to another router (router 2), it sends the packet keeping the target PC’s network address but replaces the MAC address with the second routers MAC address.
Router 2 receives the packet and looks up its routing tables to resolve the target PC’s MAC address. If the router finds the target PC within its tables the packet is sent along with the target PC’s network address and the target PC’s MAC address.
This is how a computer on one end of the world can send a message to the other end without knowing the target PC’s MAC address. If at any time a router receives a packet that it can’t resolve and doesn’t know where to forward it, the packet is simply dropped.
Optimal Path
There are many routes that a packet may take to reach its destination. It could pass through one router are undergo a dozen hops before reaching its target. The ultimate route depends on line conditions, congestion and various other factors.
Routers transmit information amongst themselves, basing their forwarding decisions on configuration tables (routing tables), that reference the network address, as opposed to switches and bridges that use the hardware address. They add an extra layer of functionality in that they cooperatively determine the best path for data packets based on line conditions and traffic congestion. Routing algorithms consider reliability, delay, bandwidth and load when determining a packets next destination or hop.
Routing Protocols
Routers can be grouped into large clusters known as autonomous systems. Groups of autonomous systems make up the Internet. Routers use the RIP and OSPF protocols to communicate within an autonomous system and another set of protocols to communicate between AS’s.
A router’s physical dimensions vary from small desktop models common for Internet sharing duties within households, to large supercomputer versions that can process 320 billion bits of information per second.
Address Translation
A computer’s true network address is its MAC address or hardware address. This series or numbers is hardwired into the computer’s network interface card and is a unique number that distinguishes it from all the other network cards in the world. Unfortunately addressing data packets using the MAC address imposes serious restrictions on network size because there is no facility to route packets using hardware addressing. IP addressing overcomes this restriction by its ability to combine computers and networks into logical groups. Because of the hierarchical structure of IP addressing each router is able to either, determine the destination computers MAC address or find a router that could help bring the packet one step closer to one that can.
There needs to be some mechanism in place to translate IP addresses to hardware addresses for an IP datagram to be transmitted and received. ARP (Address Resolution Protocol) stipulates that when a TCP/IP computer first comes on the network it is to advertise its presence by transmitting its hardware and IP addresses. The other systems then use this information to update their translation address tables (ARP Cache). If a computer is asked to send data to a destination that does not have an entry in its ARP Cache, the originating computer sends out a broadcast message (ARP Request) to every computer on the network asking the destination computer to announce itself. If the destination computer announces its presence, the ARP cache is updated and the datagram is sent on its way. If the destination computer resides on a different network or subnet, the datagram is forwarded to a router.
The data packet will contain the target computers “network” or IP address and since it doesn’t know the target computers MAC address, the first routers MAC address will be substituted instead.
If the first router(router 1) cannot resolve the network address and determines that the packet must be forwarded to another router (router 2), it sends the packet keeping the target PC’s network address but replaces the MAC address with the second routers MAC address.
Router 2 receives the packet and looks up its routing tables to resolve the target PC’s MAC address. If the router finds the target PC within its tables the packet is sent along with the target PC’s network address and the target PC’s MAC address.
This is how a computer on one end of the world can send a message to the other end without knowing the target PC’s MAC address. If at any time a router receives a packet that it can’t resolve and doesn’t know where to forward it, the packet is simply dropped.
Optimal Path
There are many routes that a packet may take to reach its destination. It could pass through one router are undergo a dozen hops before reaching its target. The ultimate route depends on line conditions, congestion and various other factors.
Routers transmit information amongst themselves, basing their forwarding decisions on configuration tables (routing tables), that reference the network address, as opposed to switches and bridges that use the hardware address. They add an extra layer of functionality in that they cooperatively determine the best path for data packets based on line conditions and traffic congestion. Routing algorithms consider reliability, delay, bandwidth and load when determining a packets next destination or hop.
Routing Protocols
Routers can be grouped into large clusters known as autonomous systems. Groups of autonomous systems make up the Internet. Routers use the RIP and OSPF protocols to communicate within an autonomous system and another set of protocols to communicate between AS’s.
A router’s physical dimensions vary from small desktop models common for Internet sharing duties within households, to large supercomputer versions that can process 320 billion bits of information per second.
Switches
Switches come in two flavors: “Cut through” which examines only the packets destination address, disregarding any errors within the packet, and “store and forward” which examines the entire packet and perform CRC error checking on the frame. Today both technologies are about the same speed. Similar to bridges, switches divide the network into collision domains and operate on layre2 of the OSI model.
Increase network efficiency
In a fast Ethernet bandwidth is shared amongst all the computers attached to a segment. Therefore on a 100 Mbps Ethernet, if a hub has 8 ports with a computer on each port, the total bandwidth is divided by 8 or 12.5 Mbps per port. If there are 40 computers per hub network performance could be seriously degraded. A switch solves this performance problem by giving each computer 100% of the bandwidth for a sort amount of time. Furthermore a switch directs the data to the target computer as opposed to the un-switched method of broadcasting data to every computer in earshot.Switches improve network performance by filtering bad packets. Problems are isolated to the errant segments and are kept from disrupting the entire network.
Electric signals weaken as they travel the length of a wire thereby limiting the distance that a good signal can travel. A further important benefit of switches is that they regenerate and re-time packets allowing each port on the switch to be treated as an individual Ethernet segment with all the configuration rules that implies.
Switches can create several dedicated connection between two computers. They can hold several open line of communication simultaneously.
Increase network efficiency
In a fast Ethernet bandwidth is shared amongst all the computers attached to a segment. Therefore on a 100 Mbps Ethernet, if a hub has 8 ports with a computer on each port, the total bandwidth is divided by 8 or 12.5 Mbps per port. If there are 40 computers per hub network performance could be seriously degraded. A switch solves this performance problem by giving each computer 100% of the bandwidth for a sort amount of time. Furthermore a switch directs the data to the target computer as opposed to the un-switched method of broadcasting data to every computer in earshot.Switches improve network performance by filtering bad packets. Problems are isolated to the errant segments and are kept from disrupting the entire network.
Electric signals weaken as they travel the length of a wire thereby limiting the distance that a good signal can travel. A further important benefit of switches is that they regenerate and re-time packets allowing each port on the switch to be treated as an individual Ethernet segment with all the configuration rules that implies.
Switches can create several dedicated connection between two computers. They can hold several open line of communication simultaneously.
