Professional Documents
Culture Documents
Become A Network Admin
Become A Network Admin
Networks can have different physical and logical topologies; physical wires can follow one
pattern while logical moves data through network in different pattern
Partial Mesh: all devices directly connected to at least two other devices
- provides strong redundancy, but not full
- not as expensive as full mesh
- multiple NICs and cable runs on each machine
- most likely found in WAN. Internet is a partial mesh.
Not all cities are connected to each other, result is do not have to have a full mesh where every
city is connected, reducing cost.
Bus topology: one of the oldest networking topologies; all nodes connect directly to the
main cable, called the bus; simple to put together,only one backbone cable; disadvantage only
one node can send a signal at a time.
To determine which node can send the signal, contention is used. Each computer on the bus
network competes to send the signal first.
Downside, if two devices listen and don’t hear a signal and send at the same time, a collision
occurs. The more active nodes, the more collisions on network.When collisions happen, signals
cancel each other out. Too many collisions can bring down the network.(network storm) This is
the idea behind the DoS attacks in network security. Recommended for networks with less than
30 nodes. Single bad node or cable can bring down the whole network. Not part of current
TIA/EIA 568-C standard for LAN networks.
Must make sure there is a terminator (terminator resistor) on both ends of the main bus cable.
Ring topology: older network; similar to bus (one backbone cable) but connected in a
circle. packets move around network in a ring; each node is given an opportunity to send a
signal, causing no contention between nodes. Heavy traffic can slow down network, but not
bring it down. Like bus, single damaged node or cable can bring down network. Also not part of
the TIA/EIA 568-C standard.
Star and Hybrid Star Topologies
Hierarchical Star Topology: most commonly used in LANs; more expensive than bus
because of the need for more cables. One damaged node or cable will not bring down the
network. All nodes connect to a central hub or switch, making it easy to troubleshoot a
hierarchical star configuration.
Disadvantage is it’s susceptible to a single point of failure. If whole network goes down, central
device is a problem, so you know it’s the central device that everything is connected to.
If only a single node is down, the node is the problem.
Biggest advantage, only topology recognized in TIA/EIA 568-C for LAN networks.
-Physical Hybrid Star Topology: network containing two or more physical topologies within
the network itself; part of network may be physical ring while another part of the network may be
a physical star
Most commonly, multiple switches within a network and each one of those switches are linked
together as a bus, from each switch there will be a physical star going out from the switch.
Result > physical bus connection our switches and logical stars contained on each switch.
-Physical-Logical Hybrid Star Topology: network physical looks one way but functions
differently
A common physical logical topology is a network that functions like a ring, but looks like a star.
Most IBM token ring networks follow this pattern; a central device that all the nodes are
connected to, but inside central device it behaves in a ring pattern.
Another physical-logical hybrid topology is a star network that uses a hub. Various nodes
connected to the hub in a physical star pattern, but inside the hub data is moved about in the
shape of a bus.
Point-to Multipoint Topologies: crossover cable connecting two different computers together
enabling the two computers to talk to each other directly without having to use a hub or switch.
-is often part of a hybrid system where point-to-point connections will often form the bus
connecting multiple connecting devices together
-point to multipoint is same, except one device connects directly to more than one other
device
Illustration shows each of the nodes are computer, but can also be switches, routers, or other
devices.
The little symbol inside each computer is a policy and security configuration. Each computer is
responsible for its own management and for management of the resources it may be connected
to.
Client Server Network Model: all devices access resources through central server
- devices needing access to network are called clients
- device/devices controlling access are called servers
- network management overseen by server
- security build around access to server
- server allows or restricts access to network resources
- server controls who is allowed to log on to network
- disadvantage, if the server goes down, no one is able to access network resulting in no
resources being accessed via network until it is back up
Six computers in network, but each computer does not have a security policy inside. Each
computer does not control security for the network, even for itself. Instead, it is controlled by the
domain controller. This is shown by the symbol for security policy located on domain controller.
In order for a desktop to get access to the network, it has to go to the domain controller and get
permission of what it can and can’t do.
Once authenticated, it can gain access to any resources on the network that the domain
controller allows.
A collision took place and a timer on each computer on the network randomly generated a
number of milliseconds and set the clock to that number of milliseconds. Each node has a
different number of milliseconds on their clock. As each node clock runs down they attempt to
send their packet and this way is hope that they will avoid additional collisions.
Carrier Sensing Media Access/Collision Avoidance (CSMA/CA): most common in
WiFi networks.
- works similar to CSMA/CD; difference is that it releases warning packet before
releasing data packet
- if other nodes hear warning packet, they won’t transmit
- once data packet heard, other nodes are able to transmit
- two warning packets transmitted at same time, this will cause a collision; collision
is handled the exact way as the CSMA/CD method.
*side note: WiFi doesn’t stand for anything in particular; inventor thought it sounded cool
Computer 1 wants to send packet to 3. Sends a warning on the network. All computers on
network hear warning, and refrain from trying to send their own warning. Once warning has
gone by and others heard it, 1 will release its data packet onto the network.
Internet is NOT the Worldwide Web, though many use it interchangeably. The Worldwide web is
just a service that runs on top of the Internet.
II NETWORK IMPLEMENTATIONS
WANs and MANs
WANs
- one large network that covers a large geographic area
- Internet - best-known WAN
- can be many smaller networks linked into one large
- called an enterprise network when owned by one organization
- use routers and switches to connect up network
MANs
- uses same technology as WANs
- covers an area of only 50 kilometers or so across
- term falling out of use; WAN used to apply to both WANs and MANs
WLANs
- is a standard LAN that uses wireless technologies
- wireless technologies commonly referred to as Wi-Fi
- Wi-Fi technologies defined by IEEE 802.11 standard
- IEEE 802.11ac- most current Wi-Fi standard
- more advanced IEEE 802.11ax standard being developed
Medianets: networks optimized for distributing large video apps and similar technologies
- Hulu, Netflix, WebEx, GoToMeeting
- uses smart bandwidth detection systems
- allows medinets to adjust to higher or lower bandwidth devices
- provides smooth video transition on any platform
OSI MODEL
Intro to OSI Model (open systems interconnection reference model)
- created as reference model and teaching aid
- not intended to reflect network architecture
- model for organizations creating new protocols
End-to-End Layers:
- transport layer and above layers not responsible for transmission between nodes
- transport and above layers responsible for source to destination transmission
- source-to-destination transmission also called end-to-end transmissions
- upper layers not concerned with underlying communications facility
Session Support:
- performs the function of allowing processes to communicate over a network
- performs security
- performs name recognition
- performs logging on
- performs other functions that are less common
Application header tells computer on the other end what applications or services running on the
OS are being used for this data.
Presentation header contains info about encryption, compression, translation or anything else
the computer on the other ends needs to know in order to identify and use the data that’s being
transmitted.
Session layer presents info needed by the session layer on the receiving end to setup the
session between the two processes that the data applies to
Transport layer has info about specific protocol or process needed by the session layer to set up
the session with.Network header provides info about the overall final destination of the data
that’s being transmitted.
Data link layer provides the address of the next node that the data needs to pass on to in order
to reach its final destination. Also includes a trailer that contains some error collection info.
All information is converted into 0s and 1s or encoded into 1s and 0s and put on the media
being used to transmit the data. As data moves through the OSI model, it becomes larger and
larger. (overhead)
When data is received, it removes a header through each layer. Telling receiving computer what
to do next with data it has received.
TCP/IP Suite
- group of protocols designed to work together to send data across a network
- named after the two major protocols in suite
- TCP: transport control protocol
- IP: internet protocol
- contains a large number of protocols that are able to carry many different network functions.
- open protocol suite, free for all to use
- new protocols can be freely developed as needed
- only protocol used by the Internet
- all existing networks use TCP/IP as main transmission protocol
TCP/IP
- created by DoD in 1970s
- reduced version of OSI model
- based on and around TCP/IP protocol suite (will match up with the different layers of TCP/IP)
- 4 layers
Application layer
Transport layer
Internet layer
Network access/network interface/link layer
- all TCP/IP protocols are located on the top 3 layers
- protocols located on bottom layer are not part of TCP/IP suite
- each layer corresponds to one or more OSI model layers
Hubs
- older technology, falling out of use in favor of switches
- work on layer 1 of OSI
- logically function as bus topology
- too many hosts can result in constant collisions, can bring down entire network
- intelligent hubs: really active hubs with additional features; some have network diagnostics
abilities; some have management abilities; other availabilities are available;
Understanding Bridges
Bridges
- device used to breakup a network into smaller segments
- is an older technology
- works on layer 2 of OSI
- can read frames to determine if they are allowed to pass
Exploring Switches
Switches
- device used to connect multiple computers
- primarily work on layer 2 of OSI
- some can work on layer 2 as well as higher layers
- when a switch can work on more than one layer, called multilayer switch
Basic Switch
- most common type
- essentially a multiport bridge
- used to separate larger networks into smaller segments
- segments called collision domains
- uses ports to set up point-to-point connections between devices connected to affected ports
- results in no collisions on network
- allows different ports to communicate at full speed
- makes it harder to listen in to traffic on the network
- commonly used to convert media from one type to another
- some switches have fiber coming into them
- but sends signals out over copper twisted pair wires
Understanding Routers
Routers
- move data around large networks like WANs
- work usually on layer 3 and 4 of OSI
- are intelligent
- make independent decisions about sending data
Numbering Notations
-notation system to avoid confusing numbering systems
-1101 can be a decimal, hexadecimal, or binary
-following applies when viewing numbers on a computer
- 1101 by itself usually indicates a decimal value
- 0x1101 indicates hexadecimal value
- 0b1101 indicates binary value
Physical Addressing
-refers to actual physical address assigned to a device
-is set at the factory; can’t do after the fact
-often referred to as MAC address
-term MAC address is relevant to Ethernet
Now has unique address for every device. Routers (round) computers have MAC addresses,
but switches (boxes) do not due to port and switch takes on whatever it’s connected to.
Routers have multiple interfaces and have multiple MAC addresses; cannot use same MAC
address twice on same network. Physical address assigned to every interface on network.
Segments are formed when 1 or more switches connect computers to a router's interface or
when two routers are connected to each other. Or just switches with no router. However without
router, cannot connect to outside network.
In above diagram, each switch along with computer and router interface become a segment.
Interface that connect router directly together become a segment. 10 different segments in
above diagram.
Logical Addressing
-all devices on network need unique logical address
-logical address is determined by network layer protocol being used
-almost all modern networks use TCP/IP
IP Addressing
-entered manually or automatically when config interface
-manual IUP addresses also called static ip addresses
-there are 2 main methods to dynamically assign IP addresses:
-dynamic host control protocol (DHCP)
-automatic private internet protocol addressing (APIPA)
DHCP: requires server set up on network and connected to segment you want to assign IP
addresses to
APIPA: only usable by Microsoft operating systems
- all addresses will fall in the range of 169.254.0.1 to 169.254.255.255
-most Microsoft operating systems failover to APIPA if no other addressing is available
-many routers and switches intended for home users use APIPA as default
0 at end of IP means network address
When network address to network becomes network or subnetwork
One network is assigned to subnetwork, unique host IP addresses can be assigned to each
interface within that segment. Once you have a subnet address, can use that for unique host
addresses. Router has physical address A, but logical address of 192.120.201.1. Computer C is
physical address, but