Lesson 4

Okay, so a bit late but I have been having a hectic week - I thought that I would sneak this in at work! Don't bust me!

So in this lesson we looked again at the planning of a site: (inside of the OSI model covering, Network, Datalink and Physical)

• Where to put the backbone (if there is a central lift shaft this is good!)
• Where to put the IT department (in the middle somewhere to reduce the length of cable between the workstations and the server)
• Decide on what medium to use for the backbone (we looked at fibre or STP dependending on the budget and how fancy you want to be)
• Decide on a topology (per floor or department in the building)
• Remember that the backbone needs to be terminated at each end!
• Decide on the minimum data speed 100/1000
• Once a decision is made on the backbone speed, everything else should be in line with this decission - of course the last section connecting to the workstations would be fine with 100 megs
• Decide on redundancy
• Server with a mirroring server
• Server connected to a switch which distributes to switches on each floor
• Switches on each floor could be connected in duplicate to the switch at the server

The switch configurations are between:

• 8, 16, 32, 64, 128 ports - decide on what is necessary, this would depend on the number of employees with an idea of possible growth and upgradeability.
• WAP - wireless access point is another switch which you could use

Remember that the software controlling the network now would be the Domain Controller (we would be using Microsoft Active Directory)
With the redundancey and the mirrioring of the servers you have two options:

1. Intrasite - means that both servers are on site so both domain controllers are mirrored on site
2. Intersite - means that there is off-site replication (domain controller on-site replicates with a DC off site to cover you in the event of fire and floods etc)

The intrasite replicates every 8 minutes, and intersite can be setup as per your requirements (best to do this at like 3 in the morning when the network is quiet. It all depends on the clients needs and the type of business.
'
Load balancing - this is when the domain controllers can replicate in both directions to share the resource when the network is under pressure (help each other out)

With the domain controllers there are two types that we looked at this weekend and they are:

1. Post 2000 - this has no hierarchy and both would be equal, thus the the DC can be primary or back-up
2. Pre 2000 - (NT server) here there was hierarchy and the DC would be primary or back-up, this was PDC or BDC (primary domain controller or back-up domain controller)

We then looked at the term collision domains, if you look at the building that we designed, each floor would be a collision domain from its switch on the floor. By doing this you minimise the collision domain thus ensureing a beter flow of data.

We looke at the Firewall - this could be the server - a Proxy server can be chosen to aid in the firewall - this is created in the domain controller server. The proxy stores the sites in cache and retains the handshakes up until the time you shut down your machine (like the proxy when it does this on your behalf!). The Proxy is the DMZ (demilitarised zone) on the domain controller. - this is a bit confusing, do some more research and draw your own pictures!
The area between the proxy adn the domain controller is called this DMZ. Proxy's can be set up to filter sites visited like porn etc (must filter Facebook!!!) The Proxy can also filter ports like a port associated with the Internet is port 80.

in the Network layer (MESH) and there are routers involved when you have a WAN remember, as soon as you have more than one network card it becomes a router!

A router changes a public IP address into a private IP address ( it translates like MODEM this can also be a gateway)

We then looked at the transmission methods:
Baseband and Broadband - baseband uses all the available bandwidth and broadband uses only sections of the bandwidth.
Bandwidth can be used as:

• Unicast - like a two way radio (simplex)
• Broadcast - like SABC (this goes to everyone - switch or hub using 802.3)
• Multicast - like M-net - only broadcast to certain specific groups

Further definitions below:

• Unicast - Unicast transmission is a method where data is transferred fro a specific source address to a specific detsination address
• Broadcast - is a method in which data goes from a source node to all other nodes on a network.
• Multicast - is a transmission method in which data is sent from a server to specific nodes that have been defined as members of a multicast group.

Access methods:
Communication method categories include transmission, media access and signalling.

Media Access methods
 Only 1 node can transmit at a time, the media access method determines whether or not a particular node can place data on the network a any given time. Ther are two categories: contention based or competitive media access. The nodes negotiate themselves for media access time with deteministic medai access a centralised device or system contriols when and for how long they can transmit.

Multiplexed Media Access
Multiplexing is a controlled media access whcih a central device combines the signals from multiple nodes and transmits the joined signal across the medium.
Signals can be multiplexed by using either TDM (time division multiplexing) or FDM ( frequency division multiplexing) the central device that the rely on is called a MUX (multiplexer) - at the other end a DEMUX or demultiplexer separates the signal again.

• TDM - divides the signal into discrete time slots
• FDM - data is sent over different frequencise or channels (put together again on the other side)

Polling
(token ring environment)
Polling is s acontrolled media access method where a central device contacts each node to see if it has data to transmit, each node has guaranteed access but time is wasted polling nodes that have nothing to transmit.

Demand Priority
This is a polling technique  where the nodes signal their state - either ready to transmit or idle not ready to an intellegent hub. The hub favours high pripority transmission requests.

Now for an important bit of revision! NBNBNB

CSMA/CD adn CSMA/CA - carrier sense multiple access with dollison detection or collision avoidance!

CSMA/CD - this is a contention based media access method, they must detect and manage the inevitable collisons that occur on the network when mulitple nodes transmit at the same time. They detect a collision and then re-transmit at a later time, these times are set at random amounts to avoid further collisions.

1. A node has to transmit
2. The node determines if the media is available
3. If it is it transmits the data
4. The node determines if a collision has occurred by detecting the fragmented data that results from the collision
5. If a collision has occurred, the node waits for the random "back-off" period which is in milliseconds and the repeats from 1 above.

CSMA/CO - this is also a contention based media access method used by 802.5 and 802.11 LAN's. In this setup the nodes can transmit whenever they have data to send, they do however takes some steps beforehand to ensure that the media is unused adn that they can send without collision.

1. A node has data to transmit
2. The node checks to see if the media is available
3. If it is available it transmits a "jam" signal, which lets the other nodes aware that it intends to send (advertises)
4. The node waits until all the other nodes should have had time to recieve the jam signal
5. The node transmits its data
6. While it is busy transmitting it monitors the media for a jam signal from other nodes on the network, if it recieves a jam signal from another node, it stops transmitting adn retries after a random delay.

Contention Domains

This is also known as a collision domain - this is where a groiup of nodes compete with each other for access to the media. So, by deviding your network in various / many contention / collision domains will reduce the possibility of collisions.

Each time that you have a DC you have a domain, each time that you have a switch you have a collision domain - think of the building we planned!
A collision / contention domain is the SAME thing.

The last topic is SIGNALING METHODS:
Analogue Signals
Analogue signals form a contiuous wave
Digital Signals
Unlike the analaogue signal wave which can have multiple values depending on the amplitude the digital wave only has two values, ones and zeros (on or off)

Terms:

• 
  
  Amplitude: - the height or depth of a wave from its midpoint to the top or the bottom.
  
  
  
• 
  
  Cycle: - One complete oscillation
  
  
  
• 
  
  Frequency: - The number of complete cycles per second (measured in hertz)
  
  
  
• 
  
  Phase: - Where the cycle begins in relation to a fixed point, so if two waves start at the same time - they are in phase - think of electricity and a video camera showing lines on a tv
  
  
  
• 
  
  Wavelength: - The distance between the successive peaks in a waveform
  
  
  

That's all for now - expect the next post next week!

Lesson 3

With this lesson we continue with Topology - but, we start to look at the LOGICAL topology now. Remember that Physical was the tangible side of things, thus the Logical is the flow of data - we cannot see this! :)

• Logical Bus Topology

In the logical bus topology the signal sent is received by all the nodes at the same time, so one signal is broadcast and all the nodes receive the signal. - this is 802.3 CSMA/CD

• Logical Ring Topology

In the logical ring topology, the signal is sent around the Ring like a dog chasing it's tail - it always travels in a clockwise direction. (a token continually travels around the ring looking for information to pick up from the network card and transport it to the required destination)

• Logical Star Topology

This is a topology which is implemented less frequently than a Ring or bus. In the star, the nodes are connected to a central device and are all on the same bus cable. The central device continually polls the nodes to determine if the need to send information. This can be more complex and a MUX (multiplexer) can be used - this will split up different sections of the network and send only to the section that requires it.

• MSAU / MAU - multi-station access unit

In this situation, the logical ring topology has a MAU whereby the signal travels in the ring inside the unit and then sends the signal down the last section of cable once it has established its location.

When you have two different standards, namely 802.3 and 802.5 in a situation where you have a MAU, for the two topologies to be able to connect, you would require some sort of "gateway" which translates the one standard into the other so that they may communicate - this is also a good way to safeguard your network (security).

In Lesson 3 we had a look at what we would need to consider should be be designing the architecture of a network in a building, in the points following we will list the necessary steps or things to consider.

• How many floors is the building? Always try to put your server room in the centre of the building to reduce cable pulls.
• Where is the lift shaft? Usually in the middle of the building - this shaft provides you with a good spot to run your cables from floor to floor - cables need to be shielded in this environment.
• decide on a topology for the network per floor
• decide on a minimum network speed - then ensure that your back-bone which runs up and down the lift shaft is of maximum speed.
• a patch panel (there would more than like be one for the telephone system) will be required on each floor to distribute that cables to each topology on each floor, this patch panel will hold the switch and the switch is connected to the server via the back-bone.
• The switches on each floor could also be MAU devices.
• Server room should be off the ground level if possible - water issues
• Your back-bone through the lift shaft must be terminated on each end.
• where the back-bone connects to the server, you need to focus on the failure points here, you could have a switch or multiple switches between the back-bone and the server or servers to lessen the chances of failure due to cable breaks.
• Server mirroring will also reduce the effects of failure, redundancy is important!

• Network Categories

LAN
Local area network - this is a self-contained network which covers a defined small area such as an office. Can be connected directly via cable or via wireless technologies

WAN

Wide area network - this is a network which spans a large and multiple geographic locations (much like a national or international company) it would connect multiple LANs.

Now this gets interesting as a WAN is the basis of the Internet as when a a device has more than one connection it becomes a Router. Thus each LAN is connected via the WAN which includes TELECOMS.

A WAN will always use a MESH topology. SO you can see how the Internet is a giant WAN connecting all the computers around the globe, A Mesh has at least 4 connections per workstation.
Lastly there are some other Networks listed below:

• MAN - metropolitan area network, imagine the city's network
• CAN - Campus area network, imagine the network that UCT might have
• PAN - Personal area network, this connects 2 or 3 workstations together, imagine the network you might have at home or in a small office.
• WPAN - wireless personal area network, this is a small network much like the PAN but uses wireless technology rather the a wired/cabled network.

That's all for now - next update after lesson 4.

Lesson 2

So, I was unable to make it to this class so there might be one or two things missing or extra as I tried to fill in the blanks.

Lesson 2 primarily focuses on the topology of a network, this can be split up into:

• Physical topology (architecture)
• Logical topology

The physical topology is how the network actually looks, how is has been laid out with cables etc.

The logical topology is the route that the data follows - this is what you can't see!

The physical topologies that we look at are:

• The Physical Bus topology

This is where nodes are arranged in a linear fashion - one after the other, each node is directly connected to the network cable (backbone) with a T-connector and then by a BNC connector directly to the node. Think of peer-peer. These types of networks are not very reliable as if there is a break in the cable, the entire network goes down, they are also typically slow and difficult to troubleshoot. The backbone must have terminators on each end so that when information is broadcast, it knows where to end! (can't use UTP in bus)

• Physical Star Topology

This type of architecture uses a central device from which all nodes branch off or connect. A device such as a HUB or Switch. When using a hub, the signal is broadcast to all, when using a switch there is some intelligence as the switch has record of where the nodes are situated on the physical topology. Thus it will be a bit faster.

• Physical Ring Topology

This is a layout whereby each node is connected to its nearest node and the signal travels like a dog chasing it's tail. Each node then amplifies the signal to the next node. The signal always travels in a clockwise direction around the ring. The nodes will have network cards with an "in' and "out" point which determines the upstream and downstream node. Failure of a node could potentially bring down the entire network.

• Physical Mesh Topology

This is when each nodes is connected to each and every other node

In this mesh topology, each node has 4 network cards in it, the mesh is very difficult to fail as it always has a redundancy to the entire network. In the Physical Mesh, the data will always follow OSPF - open, shortest path first! NB!!!

• Hybrid Topology

This is exactly as it says - it is a hybrid between two or more topologies.

The acronyms that we need to know and some other points to remember:

• T.T.L. - time to live (the amount of time given to a signal, its life)
• UTP - unshielded twist pair (cable) - Connects to a RJ45
• STP - shielded twisted pair (cable) - connects to a RJ45

• Coax - cable with a core and a protective sheath (like television cable)
• Fibre - shielded cable with a glass core
• CSMA/CD - carrier sense multiple access with collision detection (802.3) Ethernet
• CSMA/CA - carrier sense multiple access with collision avoidance (802.5)
• UTP - cannot be used in a bus topology
• In the exam, we will need to know and be able to draw the types of topologies!
• Below is some cable information: