Network Geography

PAN - Personal Area Network

Usually something within arm's reach.

Examples:

Bluetooth, USB, Firewire

LAN - Local Area Network

Usually up to 100 meters. Small office or 1 floor of an office building.

Examples:

Ethernet (802.3), Wi-Fi (802.11)

CAN - Campus Area Network

Building-centric LAN. Spans entire University. Associated with business parks and
military bases.

MAN - Metropolitan Area Network

Area of up to 25 miles. Multiple CANs can create a MAN. Associated with

distributed college campuses or distributed organizations.

WAN - Wide Area Network

Large geographic area, across a state, country, or even the world. WANs do not
have to be public.
 Personal > Local > Campus > Metropolitan > Wide

P. L. C. M. W = PLC Mechwarrior

Wired Network Topologies

Topology Types:

Phyiscal (Cabling)

Logical (Flow)

Bus Topology

Has "Terminators" at each end to end signal propogation.

Ring Topology

A ring topology using an electronic tag "talking stick" is known as a Token Ring.

FDDI Ring

Still used in Fiber networks. Mentions of "Ring" on the Net+ Exam are referring to
FDDI Ring.

Star Topology
 1 control device; is a SPOF (Single Point of Failure).

Hub-and-Spoke Topology

Full-Mesh Topology

Maximum redundancy, but not practical. Used for Nuclear C2.

Partial-Mesh Topology

Some portions are fully redundant, some partially, and other not at all. Better
redundancy of "Hub-and Spoke".

Wireless Network Topologies

Infrastructure Mode

In-line Modem and Router or some sort of similar configuration. Adds additional
security.

Ad-Hoc Mode

Decentralized, device-to-device connections.

Wireless Mesh

Interconnection of different types of nodes and media. Includes cellular, wireless,

microwave, and satellite.

7 Layers of the OSI Model

The model is not 100% accurate, but categorizes functions is some meaningful
way. Networks now operate using the TCP/IP model. OSI model is generic in
nature and works for all networks.

As information goes through different layers, we call it different names.

Physical Layer (Layer 1)

Examples:

Ethernet, Fiber, Copper, CAT5, CAT6, WiFi, ...

The Physical layer is always concerned with Bits. Layer 1 devices view network
 topology from a Physical perspective. It uses Transition Modulation - +0.x/-0.x
voltage represents a bit being transmitted, or a light being turned on and off.
Physical layer is also concerned with the wiring of the cables / the standards of
wiring. For Ethernet, there is TIA-EIA-568B and "568A.

Straight-through cable means that the cabling order on both sides of an Ethernet
cable are the same order.

Synchronization of Communications:

Asynchronous communication is like a voice-mail. It is done using a Start

and Stop bit. Synchronous communication is like a telephone call. The
route and data, respectively, is static and done at the same time. In
synchronous communication, a common time source is used.

Broadband - device bandwidth is split into different channels on the same line.

Baseband - device bandwidth is not split, "All of the signal for all of the data".

TDM (Time Division Multiplexing) - "Time is shared"; time slots are used for each
client for communication. StatTDM - dynamically allocating time slots if the time
is not being used; "If nobody is talking, then anyone is free to talk". Layer 1 devices
are "dumb" devices. What they send in is what they get out. Think of the way a hub
works.

Data-Link Layer (Layer 2)

Contains MAC (Media Access Control) Addressing. MAC addresses are 48-bit
physical addresses. They consist of 12 Hexadecimal characters, each representing 4
bits.

Layer 2 devices view and interpret the network Logically.

 LLC (Logical Link Control) is the most basic form of Flow Control & basic form of
Error Correction function. It adds checksum data, and compares the last bit for an
even or odd. Review more about this in-depth.

Synchronization is either Isochronous (common references and time slots) or

Synchronous (common reference clock but with beginning and end frames).
Asynchronous uses own clock cycles and own start and stop bits.

Examples:

NICs, Bridges, Switches, MAC Addresses

Network Layer (Layer 3)

Concerned with Routing and Logical Addressing. IPv4 and IPv6 are the most
common forms on Logical Addressing.

There are three types of switching (alternative term for routing):

Packet Switching - Data is divided into packets and sent.

Circuit Switching - Dedicated links / lines of communication. Virtual

connection; same path for the entire session of communication.

Message Switching - Similar to packet switching, but information is stored

and forwarded.

Route Discovery and Selection:

Can be Static or Dynamic; Dynamic routing is used with RIP, OSPF, and
EIGRP protocols.
 Layer 3 connection services augment Layer 2 connection services. Flow control is
used to "hold that data" or "speed it up". Packet Reordering is used since packets
can go out-of-order or follow different paths.

ICMP (Internet Control Message Protocol) is used for Error Messages, Operational
Information about an IP destination. It is not regularly used by end-user
applications, and most often used for troubleshooting. Ping and Traceroute
commands are also very common.

Examples:

Routers, Multi-layer switches, IPv4 Protocol, IPv6 Protocol, ICMP.

Transition to Mike Myers

Meet the Frame
Data is sent in discrete chunks called Packetized Data. A single frame can be up to 1500
bytes long. A frame is actually generated inside the network card. Data from applications
is sent into the network cards. NICs also ingest and destroy frames (the portions that
create an encapsulated IP packet).

The MAC Address

Identifies a frame for a specific machine on the LAN. You can use the command < ipconfig
/all > to see each network interface and the associated NIC. The CRC (Cyclic Redundancy
Check) is used to ensure the data being sent is not bad.

Broadcast vs. Unicast

Broadcast frames contain the destination MAC address of FF:FF:FF:FF:FF:FF. Every
computer on the broadcast domain will recieve and ingest the frame that is sent over the
LAN.

Introduction to IP Addressing
The predominant version of addressing used today is IP addressing, either IPv4 or IPv6. IP
packets sit within frames. The default gateway is the connection to the router itself. A
router's routing table maps where to send data. IP packets are always encapsulated in
frames, though the L2 protocol may change / data used for encapsulation may change.
Packets and Ports
The first 1024 port numbers are known as "well-known" ports. The destination port in an
IP packet identifies, in general, which application the data is intended to be sent to. TCP is
the main protocol that we use on the internet. It is connection oriented and uses a
sequencing number with it to determine order at end destination. UDP is another protocol
and is not connection oriented.
Coaxial Cabling
RG means Radio Grade. Different coaxial cables have different resistance, which is
measured in Ohms. One of the oldest types is RG-58 (50 ohm) and uses the BNC
connector. The BNC connector has a rotate lock and is not threaded. It is used on cable
modems.
RG-59 (75 ohm) uses a TV cable threaded F-type connector. RG-59 is not very robust.

RG-6 (75 ohm) is the most common cable type for cable modems. It also uses an F-type
connector and is more robust than RG-59.

Twisted Pair Cabling

Twisted pair was originally invented and used for telephone systems. It is now the most
common type of cabling that we see on networks. It usually has 4-pairs of cable, for a
total of 8 individual strands. UTP is Unshielded Twisted Pair cabling. It is very cheap but
is prone to interference. Keep it away from electrical motors, heat, and magnetism. It uses
and RJ-45 connector.

There are two common standards for Twisted Pair Cabling - TIA-EIA-568A and "-568B.
568B is most commonly used. To swap between standards, swap the greens and oranges.
CAT Ratings
Represent the speed and distance for twisted-pair cabling. Differs in number of twists per
inch, and has nothing to do with the number of pairs. CAT 7 is Shielded Twisted Pair. CAT
3 is 10Mbps; is obsolete. CAT 5 is 100Mbps for 100 meters. CAT 5E is an upgrade from
CAT 5 and offers 100Mbps for 100meters. CAT 6 is 1Gbps for 100 meters. CAT6A is 10Gbps
for 100 meters. CAT 7 is 10Gbps for 100meters but uses a different connector than CAT 3 -
CAT 6A.

Fiber-optic Cabling

Types:

Multimode - Designed to propogate light, uses LEDs, and has two connectors. Is
represented by Orange and Aqua (special) jacketing. Two connectors is equivalent
to "duplex" communication.
 Singlemode - Designed to propogate laser light, uses lasers, travels long distances,
and has two connectors. Is represented by yellow jacketing.

Connectors:

ST Connector - one of the earliest types, and kind-of looks like a BNC connector.

SC Connector - square, and uses a "punch-in, pull-out" method of connection.

FC Connector - screw in, similar to a cable television connector.

LC Connector - new, uses 2 fibers, but is smaller and has a "2-in-1" connector.
 MT-RJ - new, very small, also uses a "2-in-1"

connector.

Contact Polishing - Physical Contact method uses slightly rounded edges. Ultra Physical
Contact method uses even more rounded edges and loses less light than normal PC
method. Angled Physical Contact method uses a 70 degree angle and is the best contact
method.

Fire Ratings
Plenum - space between drop & actual ceiling or raised floor. Plenum rating is the highest
fire rating and fire resistance. It costs 2 to 3 times the amount as PVC. Raiser - space
between floors in a building. It does not have quite the same fire resistance as plenum.
PVC - non-plenum, no fire protection at all but still very resistant to fire. Refer to the
cabling box to determine the fire rating.

Legacy Network Connection

Serial Ports - low speed, two-way ports. Uses a language called RS-232. Either uses DB-9
or DB-25 connectors. Parallel Port - still uses a DB connector, but is female. It is usually
used for printers but also has been used in the past for primitive bus-type connections.
There is one exception to legacy connections. Many routers use "rollover" / "yost" cables
that provide a serial connection from DB-9 to RJ-45.

What is Ethernet
Ethernet - created from a standard called IEEE 802.3, which details how to make a
network called "Ethernet". 802.3 has gone through many changes since it was created.
The basic frame has not changed very much at all in the last 20 years. This allows for a lot
of compatabity options, both forward and backward. The CRC for Ethernet is called the
Frame Check Sequence.

Standard Nomenclature:

The Ethernet Frame

Jumbo Frame: 9000-bytes, but only used in high-speed, reliable networks. MTU is
Maximum Transmission Unit.

Early Ethernet

Uses CSMA/CD - Carrier Sense Multiple Access / Collision Detection - Carrier Sense
(before talking, listen, then talk), Multiple Access (multiple computers), Collision
Detection (two cannot talk at the same time, but electrically it's easy to determine
collision, uses "65,000 side dice" to create delay - client side).

10Base2 - 200 meter segments (185 actual), poor man's 10Base5.

Daddy of Ethernet
10BaseT. Token Ring was a product of IBM; it didn't use a Bus, but rather a Box (called a
Hub). Token Ring was proprietary, was expensive but robust. To combat Token Ring,
Ethernet guys condensed the Bus into a Box (hub). 10BaseT - 100meters, 10Mbps. This
was the cornerstone topology for all future internet (?) started there. No more than 1024
nodes per hub. CAT3 cables or later are required though for 10BaseT.

Terminating Twisted Pair

RJ-45 connectors are actually called an 8P8C connection.

Hubs vs. Switches

Switches have table of MAC addresses to reduce collision domain, forwards frames to
specific ports depending on host MAC address. The exception is a "Broadcast", a specific
MAC address (FF:FF:FF:FF:FF:FF). Sent out on all ports. Collision Domains are associated
with Hubs. Broadcast Domains are associated with Switches.

100BaseT
100Mbps Ethernet. Everyone is full-duplex today.

100BaseT4

100Mbps, 1024 nodes, 100m, CAT3 (all 4 pairs used).

100BaseTX

100Mbps, 1024 nodes, 100m, CAT5E (2 pairs used).

Previous were competing standards, 100BaseTX won and is now called 100BaseT. Offers
100Mbps, 100meters. 100BaseFX was next, offered 100Mbps, 1024 nodes, multi-mode
fiber, 2km range if using fiber.

Connecting Switches
Todays switches are auto-sensing on the ports, and can be connected usually from any
port using either a straight-through or crossover cable, though straight-through is most
often used. There may be older switches that use an "uplink" port and "uplink" button to
designate and use a port for switch to switch connection.

Gigabit Ethernet and 10Gigabit Ethernet

All are "1000Base<something>", and contain a bunch of strange standards to work with
Ethernet (normal) and SONET (backbone of Internet).

10GBaseSR, LR, and ER have <X>W equivalents: 10GBaseSW, LW, EW. These work with
SONET and have the same specs as the normal versions.

Transceivers
MSA = Multi-source Agreement = Cisco and everyone else working together to create a
standard. Transceivers are pluggable devices that allow people to keep one switch and use
fiber standards.

GBIC is Gigabit Interface Connector - is for ST and SC connectors.

SFP is Small Form Factor Pluggable - is designed for small form factor fiber connections -
LC connections.

SFP+ is an improved version of SFP.

QSFP is Quad Small Form Factor and is the "bleeding-edge". Used for 40Gbps Ethernet.
With fiber optic, we always have "duplex", so there is a send a receive cable.

BiDi is Bi-Directional which means that a single, single-mode fiber can use 1 line for send
and receive using two different wave-lengths of laser light. Two different transcievers are
used on two different devices to send two different wave-lengths over the same cable.

Connecting Ethernet Scenarios

Mixmatched switches can also create problems. If one is running at 10Gbps Ethernet, and
the other at 100BaseT, and the ports are NOT auto-sensing, there may be no
communication.
Introduction to Structured Cabling

Terminating Structured Cabling

Patch panel uses "110 punchdowns" to secure individual cables. Patch panel also has CAT
ratings like cable. To install the cables, line up them up in order and use the punchdown
tool. RJ-45 punchdowns for wall-outlets work the same way; put the correct color in the
correct hole.

Equipment Room
The equipment room is commonly known as the MDF (Main Distribution Frame) or the
IDF (Intermediate Distribution Frame). Equipment rooms always have a 19-inch
equipment rack. The "demarc" is the point of seperation between equipment owned by the
telephone company and the individual or company being provided service. The
demarcation point could vary depending on if it is to a house, or an office building. The
demarcation point may even be from equipment installed by telecom in an equipment rack
to a owned router.
Equipment racks usually consist of one or more of the following: a patch panel, switch,
router, and server(s). Horizontal runs lead out from the patch panel and into offices.
Equipment racks are measured in "U's", aka. "Units". Each "U" is 1 3/4 inches.

Alternate Distribution Panels

An old type of distribution panel is the 66-punchdown block - the first type of patch panel
back in "the old days". It was primarily used for phones. A distribution panel also exists
for fiber optic cabling - known as a FDP or Fiber Distribution Panel.

Testing Cable
Testers include a local and remote piece of equipment to work together. Wire Mapping is
checking if all of the wires are punched into the correct locations on both ends. Continuity
is checking if both ends (regarless of wiring) are connected and do not have breaks. To
test if distance is an issue, use a TDR (Time Domain Reflectometer). Horizontal runs for
Ethernet must be < 90 meters.
Optical TDRs exist as well for fiber.

Troubleshooting Structured Cabling

There are three areas to be concerned with: the Equipment Room, Horizontal Runs, and
Work Area.

First, look client-side for network configuration issues. Check the Network and Sharing
Center in Control Panel. Check link lights on the host and then on the switch.

Check the host's network card an IP configuration. You can also use a loopback adapter
(and ping 127.0.0.1) to check network adapter functionality. A loopback adapter is a RJ-45
with pin 1 going to 3 and pin 2 going to 6.

Next, check the patch cable in the work area. Check the plug in the wall, as it is often
subject to damage. You may need to replace the plug.

Look at the Equipment Room / MDF / IDF. Check electricity (using a voltmeter) going into
the equipment. Another solution is to use a voltage monitor to monitor voltage trends. A
UPS, or multiple UPS may be a solution. Use temperature / humidity monitors.

Horizontal runs are very unlikely to have problems, since cable installers should have
certified the run during installation. Use a TDR from a wall outlet to the patch panel to
test for breaks. Check for interference from other devices (possibly HVAC power, etc.).

Multimode (LED) fiber is sometimes affected by Modal Distortion.

Using a Toner and Probe

Also known as a "fox and hound" and used to find cables.

Wired Connection Scenarios

Two different categories of problems: slow or poor connectivity, and no connectivity.

For slow speed:

Attenuation - signal degrades over distance. Make sure runs are less than 100 meters
upon install. Jitter - packets dropping. VOIP and video are highly affected. Find the
bottle-neck (speed mismatch?). Incorrect cable type - use the best cable. Don't use old
patch cables, since the cables won't support higher speeds. The "speed lights" on the
switch - showing that the switch is trying to negotiate a higher speed, will show
indicators.

For no connectivity:

Bad ports on switches. Transient electrical problems (low voltage) may cause that. Use a
different port. A bad port on a switch == replace the switch. Another problem: TX/RX
reverse. Use a straight-through / crossover cable where appropriate. Bent pins - switches
can be abused, just replace the switch. Opens / shorts - don't spontaneously happen. Two
pins may be touching at termination. Replace the cable.