Transcript Document

Chapter 3: LAN Protocols
Legacy Protocols
A legacy protocol was widely used in the past, but are rarely
implemented now. You may encounter legacy protocols on
older networks.
Appletalk. Used by apple computers in the 1980s and 1990s.
Modern apple networks use TCP/IP.
DLC Protocol. Used by Hewlett-Packard, often for printers. Also
known as JetDirect. HP printers use TCP/IP now.
NetBEUI. Used by Microsoft until the release of Windows 2000.
IPX/SPX. Novell’s legacy protocol. Modern Novell networks use
TCP/IP.
TCP/IP
Stands for Transmission Control Protocol/Internet Protocol.
• Core protocol of the Internet since 1983.
• In use on almost all LANs today.
• IP is an OSI Layer 3 Protocol.
• TCP is an OSI Layer 4 protocol.
• There are two IP standards–IPv4 and IPv6.
IPv4
209.46.18.195
11010001.00101110.00010010.11000011
• In common use today on the Internet and LANs. Packet Header varies in size
• Uses 32-bit address as shown above in blue or 2^32
• When represented in decimal form, an IP address has four numbers, one for
each byte. This notation is dotted quad and takes the form shown above in
red. The decimal value of each quad is between 0 and 255.
• Certain address spaces are reserved for private and multicast networks.
These addresses can not be used on the Internet, but can be used on LANs.
• Private IP address space is most commonly used on LANs. Private address
space includes the following ranges.
10.0.0.0 to 10.255.255.255
Class A
172.16.0.0 to 172.31.255.255
Class B
192.168.0.0 to 192.168.255.255
Class C
IPv6
bits
16
16
16
16
16
16
16
16 = 128
IPv6 2001:0db8:85a3:08d3:1319:8a2e:0370:7344
• In limited use today, is likely to be in common use by the end of the
decade. Being tested on Internet II
• Uses a 128-bit address, represented as a 32-digit hexadecimal address.
Normally written as eight groups of 4 hex digits as shown above in red.
• Will allow every network device in the world to have a unique address.
• Supported by modern operating systems.
• Different IPv6 forms of expression
1080:0000:0000:0000:0000:7435:192.168.100.1
1080:0:0:0:0:7435:192.168.100.1
1080:0:7435:192.168.100.1
1080::7435:192.168.100.1
IP Version 6
• The next generation of the IP protocol is IPv6. 2^128
• 340 undecillion or 340 trillion, trillion, trillion addresses
• It uses a fixed packet header size of 40 bytes so that
information always appears in the same place.
Goals of IPv6
• To provide for transition from IPv4
• Simplify the header fields of IP
• Provide for authentication and privacy
• To expand routing capabilities
• To expand addressing capabilities
• To expand quality of service capabilities
• To improve support for options
IPv6 IP Header
Subnet Masks
255.255.240.0
11111111.11111111.11110000.00000000
• Like an IPv4 address, a 32-bit number.
• Used with IPv4 addresses to logically segment networks.
• A host uses its IP address and the subnet mask to determine
which addresses are on the local network and which are on
remote networks.
• Traffic destined for hosts on the local network is sent
directly to that host.
• Traffic destined for remote networks is sent to the router.
Network Address Translation
• Where one public IP address (one that is unique to the
Internet) is shared by hosts on the private network.
• Hosts on the Internet can not initiate contact with a host
on the private network.
• Hosts on the private network can initiate contact with
hosts on the Internet.
• Once contact is established, bi-directional
communication is possible.
Address Assignment
Addresses must be unique to the network.
o Two hosts on the Internet cannot have the same IP
address.
o Two hosts on an organization’s private network cannot
have the same IP address.
o Two hosts on different organizations private networks
can have the same IP address.
DHCP Address Assignment
• Addresses can be assigned manually or dynamically.
• DHCP is commonly used to assign TCP/IP addresses
automatically.
o Computer boots up and is assigned TCP/IP
configuration via network.
o Addresses can be assigned on a first come, first serve
basis from a pool or reserved on the basis of MAC
address.
Dynamic Host Configuration Protocol (DHCP)
Bootstrap Protocol (BOOTP)
• DHCP assigns addresses from a poll, then removes it
from pool
– Host sends DHCPDISCOVER message on local IP subnet
to find the DHCP server, using IP broadcast address
– DHCP server response with DHCPOFFER message
– Host sends DHCPREQUEST message to identify the server
to be used
– Server response with DHCPACK message with the
assigned IP for client
• Host sends on port 67 UDP
• Server sends on port 68 UDP
• Address can be reserved for a specific MAC
• DHCP Relay Agents can help cross subnets for server
Dynamic Host Configuration Protocol (DHCP)
Bootstrap Protocol (BOOTP)
• Parameters a DHCP can automatically set
– IP address
– Subnet mask
– Gateway (router) address
– DNS address
– WINS address
– Wins client mode
• BOOTP diskless operating systems, automatically
configure host during bootup on a TCP/IP network
DNS (Domain Name System)
• Used to translate friendly names such as www.emcp.com
into IP Addresses such as 209.46.18.195.
• DNS is distributed. No single server hosts all DNS records.
• Records are segmented into zones. A zone is a common
namespace.
• DNS servers that host zones near the top of the DNS
hierarchy can refer requests to DNS servers that host zones
towards the bottom of the DNS hierarchy.
DNS Addresses
• DNS addresses, also known as Fully Qualified Domain
Name (FQDN), are a collection of zone information
proceeded by a host name.
• Each element is separated by a period.
• A DNS address is read from back to front or right to left.
Host name
Top level
domain
library.unimelb.edu .au
Organization
domain name
Country Code
• au, edu, and unimelb are all separate zones, hosted on separate
DNS servers. Host name library is part of the unimelb zone.
Local DNS Servers
• Almost all LANs have a local DNS server.
• Clients on the LAN address all DNS requests to the local DNS
server.
• The local DNS server either returns the answer to the request from
its own database, or it will query other DNS servers to locate the
answer.
• In the past, DNS information was entered manually by
administrators.
• Today, many DNS servers can be automatically updated, so that
hosts that have different IP addresses can be easily contacted via
DNS name.
DNS Resolution
DNS client host1.emcp.com queries its preferred DNS server. The DNS
server in turn queries a series of DNS servers, beginning at the top of the
DNS hierarchy until it returns a result from the server that holds the zone
that the target host is located in.
DNS Name Space
Hierarchical Structure of DNS
• Translates FQDN to IP
• Root is the top of the tree (root domain) shown as . Period
• Top level domains – indicate countries, regions, org type
– 2 letters for countries (US, UK FR, CA)
– Countries sell their domain names, like Tuvalu (TV)
– 3 letters indicate type of org (.com, .edu, .pro)
• Second level domains – variable length names register to
individual or organization
– Microsoft.com, cisco.com, sc4.edu, army.mil (parent domains)
• Sub-domain names – department or geographical location
– Support, sales, training, south, west (child domains)
• Host domain – name assigned to a specific computer, this
identifies the TCP/IP host, is seen as a leaf of the tree
– Multiple host names can be associated with the same IP, but only one
host name can be given to a computer
DNS
• NetBIOS names go to Wins service, sends back IP
• DNS – FQDN are no more than 255 characters long
• IF FQDN name is requested to a DNS service, it will
return its IP
• DNS clients are resolvers
• DNS Servers are name servers
• Host files were first used, became unmanageable
• Recursive query – must have good answer or error
• Iterative query – gives a best answer, it’s here or
here is the best chance place to look
A DNS Client Will Use a Recursive Query With the Preferred Server to
Find an IP Address. While the Preferred Server Will Typically Use an
Iterative Query to Discover the IP Address
Name and Address Resolution Done on an IP Network
Sample HOST File Works with DNS server.
Sample LMHOSTS File Works with WINS server.
5-4-3 Rule for an Ethernet Coaxial Network
5-4-3 Rule
A consideration in setting up a tree
topology using Ethernet protocol is
the 5-4-3 rule. One aspect of the
Ethernet protocol requires that a
signal sent out on the network cable
reach every part of the network
within a specified length of time.
Each concentrator or repeater that a
signal goes through adds a small
amount of delay time. This leads to
the rule that between any two
nodes on the network there can
only be a maximum of 5
segments, connected through 4
repeaters/concentrators. In
addition, only 3 of the segments
may be populated (trunk) segments
if they are made of coaxial cable. A
populated segment is one which has
one or more nodes attached to it .
5-4-3 Rule 10Base-T
• What is the 5-4-3 Rule?
The 5-4-3 rule is a design guideline for 10baseT Ethernet Networks that
make use of only hubs/repeaters and do not contain bridges, switches or
routers, these devices negate the rule.
• What does the rule state?
For an Ethernet LAN of any size to operate the 5-4-3 rule must apply with
regards to hubs. There may be a maximum of 5 segments between two hosts
in a network, and there may be at most 4 hubs between these hosts and
finally there may only be users on 3 of the segments.
• What are hosts?
Hosts may be servers, workstations or printers.
• This rule does not apply to other network protocols or Ethernet networks
where all fiber optic cabling or a combination of a fiber backbone with UTP
cabling is used. If there is a combination of fiber optic backbone and UTP
cabling, the rule is simply translated to 7-6-5 rule.
10BaseT Network
5-4-3 Rule for an Ethernet UTP Network
Figure above shows the limits of the 5-4-3 rule where there are
4 hubs and 5 segments between the workstations on the left and
the workstations on the right.
If a host was attached to the top hub and another host attached to the bottom
hub? The 5-4-3 Rule would be violated, since there would be 6 hubs and 7
segments between 2 hosts.
How might I fix this violation?
Add a switch to the stack and modify how each hub is attached
to the stack of hubs as shown in Fig. 2.
By inserting a switch into the stack as shown above any host to host
communication will not violate the 5-4-3 rule regardless of where they are
attached. Remember that the switch negates the 5-4-3 rule since there are not
any more than 4 hubs/repeaters or 5 segments between any host attached with
out passing through the switch. This is only one possible solution, there are
many more.
5-4-3 Rule for an Ethernet UTP/Coaxial Network mixed.
Coaxial
UTP