|
|
Networking Technologies
Chapter 2: Lower Layer Protocols
Objectives:
Chapter 2 teaches about actual IEEE specifications for network cabling
and the protocols used on it. The objectives important to this chapter
are on page 2-1:
- Identify the function of the IEEE 802.x standards in computer
networks.
- Identify the media access procedures, transmission media, connectivity
devices, and basic design rules of the IEEE 802.3 standard and
Ethernet.
- Identify the media access procedures, transmission media, connectivity
devices, and basic design rules of the IEEE 802.3u standard and
Fast Ethernet.
- Identify the media access procedures, transmission media, connectivity
devices, and basic design rules of the IEEE 802.5 standard and
Token Ring.
- Identify the media access procedures, transmission media, connectivity
devices, and basic design rules of FDDI.
- Identify the Wide Area Network (WAN) protocols SLIP,
PPP, X.25, frame relay, ISDN, and ATM.
Concepts:
The Institute of Electrical and Electronic Engineers (IEEE) is
a standards organization that has specified many of the rules used in
building networks. Their standards are often referred to by number. In
this chapter, we discuss the IEEE 802.x standards
(there are about a dozen, currently).
LAN protocols that support the IEEE 802 standards can be referred to
as 802.x protocols. Note on the graphic on page 2-2 that most of
the 802 standards appear on the first and second layers of the ISO model,
because they are standards for physical connections. Page 2-3 begins a
list of the 802.x protocols:
- 802.1 - standards for communication to take place. This standard
maps to the first four layers of the ISO model.
- 802.2 - specifies the use of headers and frames, supporting
the LLC sublayer of the Data-Link layer.
- 802.3 - specifies the CSMA/CD access method, so this
is often thought of as the Ethernet standard. It was based
on Ethernet, which can be thought of as one implementation of the
standard.
This standard fits on the Physical layer and the MAC sublayer
of the Data-Link layer. You should be aware of the parsing of
the names of LAN types using this standard. For instance, 10BASE5 means 10
Mbps, baseband transmission and a limit of 500 meters
per segment (thick Ethernet). 10BASET means 10 Mbps, baseband
transmission and using UTP cable. (Parse: to resolve into component
parts. This is a word commonly used in English grammar.)
- 802.4 - specifies a token passing system on a physical bus
(token bus) which could be baseband or broadband, coax or fiber
optic.
- 802.5 - specifies a token passing system based on IBM's token
ring standard. IBM's standard specifies a physical ring, but 802.5
does not, so we often see physical stars that are logical rings by this
standard.
- 802.6 - Distributed Queue Dual Bus (DQDB) allows both
synchronous and asynchronous traffic. You should be aware that is supports
data, voice and video traffic.
- 802.7 - specifies how broadband transmission works.
- 802.8 - specifies how fiber optic media work.
- 802.9 - Isochronous Ethernet, a variation of Ethernet
that supports voice and data
- 802.10 - specifies methods and services for encryption
- 802.11 - specifies how wireless LANs work, like spread
spectrum and infrared
- 802.12 - 100VG-AnyLAN, is a standard that may authorize
a hub to decide which of two contenders for the bandwidth should
have it, based on priority.
The 802.3 standard covers a variety of physical implementations.
Page 2-6 offers a table of 802.3 standards, each of which defines a different
kind of LAN. Although most these would be called Ethernets in common discussions,
note that only the first column is called Ethernet in the chart, and that
it is a repeat of the 10BASE5 standard.
In the chart, look at the names and review the characteristics that
the name reflects: the first part of a name like 1BASE5 is the data
rate
(1 Mbps), the word BASE or BROAD refers to the kind of transmission
used, and the last character gives you a clue about segment length.
The length designator is the least standardized: 10BASE2 means
about 200 yards (185 meters), 10BASE5 means about 500
meters and 1BASE5 means about 250 meters. Well,
you probably won't want to use it anyway. Only 1 Megabit per second?
Page 2-7 offers more detail on how contention systems work. Note the
new fact, that if a collision is detected, a jam signal is sent.
This is the signal for stations that have just transmitted to wait a short,
random while and try again.
The text continues with more details on actual media usage under the
802.3 standard. A nice picture of a vampire tap, used with thick
coaxial cable,
appears on page 2-9. An adapter is attached to the NIC on the
PC, a drop cable runs from the adapter to the vampire tap,
which acts as the actual transceiver, and the tap bites into
the thick coax
(also called Ether Hose). Notes about a 10BASE5 LAN:
- The NIC has to be told (software or dipswitch) to use the tap as
a transceiver
- Each end of the coax segment needs a 50 ohm terminator (it
does not go on the drop cable). One end of each segment
must be grounded.
- This is 10BASE5, so the segment length is 500 meters
- No more than 100 devices per segment. Note that this includes
the repeaters you use, which will count on both segments they connect.
- Coaxial cable must follow the 5-4-3 rule. In a given
network, you can have up to 5 segments, connected by 4 repeaters,
and only 3 of the segments can have nodes on them. This seems to mean
that two of the segments can be nothing more than extension cords.
- Taps must be separated by at least 2.5 meters.
The coaxial cable is usually marked every 2.5 meters, so you know
where
you can run the next tap.
- Drop cables can be no longer than 5 meters.
- The cable is stiff, hard to turn around corners, and
hard to change.
Thin coaxial cable is used in 10BASE2 networks. The picture
on page 2-10 is pretty good, but there is an error in it. Can
you tell what it is? Check the notes below:
- This method uses the transceiver on the NIC, which is what
you would expect.
- Each end of the segment must have a 50 ohm terminator,
as above. As before, one end of the segment must be grounded.
The picture shows a ground on both ends, and is therefore wrong.
- 10BASE2 means the segment cannot exceed 185 meters
(about 200 yards).
- You only get 30 nodes per segment, including repeaters.
- Thin coax (RG-58 A/U or C/U) is used, normally with T-connectors.
- The 5-4-3 Rule applies.
- Nodes can be closer: T-connectors can be as close together as .5
meter (half a meter).
- It may appear to users that there are two connections
to the PC, because the coax must come to each PC, connect to the T-connector,
then leave the PC to go to the next node. You cannot avoid this by using
drop cables, you must use the actual coaxial bus.
- Like thick coax, this can break and suffer from bends.
Unshielded Twisted Pair cabling is the most popular medium.
Note that the text says it can be used with 10BASE5 as well as 10BASET.
Observations about UTP and 10BASET:
- The maximum segment length of 100 meters is the length of
a UTP drop from the node to the concentrator. In this sense, you
only have one
node per segment.
- Maximum number of nodes per LAN: 1024.
- Maximum number of repeaters (hubs or switches) between any two workstations:
4. This is effectively a 5-4 rule, if we consider the cables
to the nodes as segments.
- The picture on page 2-12 shows a possible configuration for a LAN.
Note the terms used: a patch cable goes to a node, a drop
cable goes to a concentrating device.
The text repeats the 5-4-3 rule for coaxial cable. Know it.
Remember that I said there were "about" a dozen 802.x standards? The
next one is 802.3u (the "u" stands for update). It is called Fast
Ethernet and comes in three types: 100BASE-TX, 100BASE-T4 and 100BASE-FX.
802.3u specifies a physical star (unlike some other 802.3 standards) and
a logical bus (like all 802.3 standards).
Each of the variants of 802.3u specifies a different medium. The text
tells us this is referred to as being a Media Independent Interface
(MII):
- 100BASE-TX
- uses two pairs of Category 5 Unshielded Twisted-Pair (UTP)
or Category 1 Shielded Twisted-Pair (STP) cabling.
- 100 Mbps speed
- for UTP, use standard RJ-45 connectors.
- for STP, use DB-9 connectors
- maximum segment length is 100 meters.
- 100BASE-T4
- uses four pairs of Category 3, 4, or 5 Unshielded Twisted
Pair (UTP)
- 100 Mbps speed
- uses RJ-45 connectors
- maximum segment length is 100 meters
- 100BASE-FX
- uses two-strand 62.5/125 micron multi-mode or single-mode
fiber media
- maximum segment length is 412 meters for half-duplex, multi-mode
fiber.
- maximum segment length is 10,000 meters for full-duplex, single-mode
fiber.
- uses media interface connectors (MIC) or subscriber
connectors (SC) specified by ANSI FDDI
Auto Negotiation means that 802.3u NICs can operate in a mixed
network (10 Mbps and 100 Mbps), and will use the data rate available
on the
LAN. The CSMA/CD method for Media Access is used for all
three types.
The discussion of Fast Ethernet ends on page 2-17, with the statements
about fast repeaters. Be aware that there are two classes of
fast repeaters: Class I and Class II. The 802.3u standard
allows one Class I repeater or two Class II repeaters to be used in
a single
collision domain. A collision domain is defined in the next chapter
as a portion of the network in which devices contend for access.
The chapter continues with a discussion of the 802.5 standard
for token rings. Remember that IBM invented it, and the IEEE refined
the definition. The picture on page 2-18 makes the concept of a star-wired
ring clearer. The standard specifies that workstations are star-wired
to Multi-Station Access Units (MSAUs) instead of standard hubs.
The MSAUs are connected together using Ring-In (RI) and Ring-Out
(RO) ports, creating a ring. When connecting two MSAUs, the Ring-In
port of one MSAU must be connected to the Ring-Out port
of the other MSAU, and vice versa. The principle is extended to
other MSAUs if the network requires more. In this way, star clusters are
connected into a logical ring.
Notes about 802.5 Token Rings:
- Either STP or UTP may be used in token rings.
- Devices connected to a port of an MSAU are called lobes, instead
of nodes
- MSAUs are connected to each other with adapter cables or patch
cables (if only a short run)
- Workstations are connected to MSAUs with patch cables
- Since many types of cable can be used, the cable distance allowed
between units varies with the type of cable.
- A token is actually a small frame or packet.
It is passed to the next station allowed to transmit by the last station
allowed to transmit.
- One station in a ring acts as the active monitor. Its task
is to remove frames from the ring that continue to circulate without
being removed by a receiver.
- The concept of beaconing is discussed. As an
example, assume there is a break in the cable between nodes 1 and
2.
Each node in a ring expects to receive signals from its NAUN
(Nearest Active Upstream Neighbor). Since the cable is broken, node
2 is not receiving from node 1. Node 2 begins beaconing, sending
a message to the ring that there is a problem. The message
includes its address, the NAUN's address, and the type of problem.
All stations
noticing the problem will beacon as well, but they will stop as soon
as they receive a packet from an upstream neighbor. The network
will
try to work around such breaks with reconfiguration.
The next topic is FDDI, Fiber Distributed Data Interface, a fiber optic
ring standard. This is an ANSI standard, not an IEEE standard, but it
makes use of the 802.2 and 802.5
standards. It is very fast, and has high capacity, making it useful for
three main applications:
- Backbones - connections to other networks that need to be
fast and wide
- Computer room networks - fast connections between critical
devices
- High data rate LANs - connections for users of data intensive
applications like CAD
Consider the fault tolerant advantage
of FDDI: it uses two rings that are counter rotating. This means
that traffic travels clockwise on one ring, and counterclockwise on
the other, making reconfiguration simple. If a break occurs
between two workstations, the rings cross over at those workstations,
turning
the two rings into one loop.
 Imagine
a doughnut. One cable runs around the outside of the
doughnut, and another runs around the inside of the doughnut. That's the
FDDI ring. Take a bite of the doughnut, all the way to
the center. That's the broken line. Now imagine the cable that runs around
the outside of the doughnut turning inside at both sides of the bite,
connecting with the inside cable, and completing the loop. This is what
happens in a broken FDDI ring. Specific factors for FDDI:
- No more than 1000 stations per ring
- No more than 200 kilometers of cable per ring
- The above numbers should be cut in half for fault tolerance
(anticipating a break)
- Multi-mode 62.5/125 micron fiber optic cable is standard
- Repeaters are required every 2 kilometers
- Class A stations are connected to both rings, Class B
stations are connected to only one ring
FDDI's token method is a bit different. The station with the token,
A, transmits its message, and tacks the token on the end of it. The next
station, B, would remove the message if it was the
recipient,
and send it on if not. If sending it on, B could also tack messages onto
the original, if it had messages to send. In this way, it is not necessary
for all stations to wait until they receive the token to send traffic.
Next, the topic of protocols used on Wide Area Networks (WANs)
comes up. You will want to be familiar with the list of protocols used
on WANs:
|
