Transcript Network Management Session 1 Network Basics

COMP3122
Network Management
Richard Henson
April 2011
Week 10 – Connectivity Issues

Learning Objectives:
– Explain the meaning of “connectivity”
– Explain the similarities and differences in
practical implementations that deliver clientserver functionality to the desktop
– identify the software and hardware
requirements for data one type of network to
be forwarded onto another
Evolution of “connectivity”

“The sum total of how well hardware &
software integrate to provide smooth
passage of data between programs on
different machines”

Now taken for granted; not always the
case
– HUGE thank you to OSI (open Systems
Interconnect)
Proprietary Systems
& “lock in”

Poor connectivity was once the norm…
– In the mainframe days, not necessarily considered to be a
problem by companies such as IBM and ICL
– “locked” customers into their products

This approach… not popular with customers
– want(ed) to be able to buy freely

Also a barrier to communication…
– between companies
– sometimes between different parts of the same company
» because they were using incompatible systems to please local
governments, etc…!
More about Open
Systems

Back in mid-1970s ISO (International Standards
Organisation) aware that the basic infrastructure
for global digital communications was rapidly
emerging in the form of the Internet
– decided that the then current proprietary isolationist
stance was not condusive to the growth of effective
digital data communications on a world-wide basis
– agreed that what was needed was open systems
Open Systems



Definition: a computer system that is “open” for
the purpose of information exchange
Open systems are therefore not restricted to one
particular manufacturers own system of
communicating
Open systems should provide the ability to:
– interchange applications and data
– between systems with different underlying hardware
and software
Open Systems Interconnect

Historic OSI meeting in Geneva…
– bold aims to produce a set of communication
protocols that would allow any system to exchange
information with any other system
– model for connectivity already out there with open
source Internet protocols developed through RFC

BUT so many proprietary models to also be
represented…
– clearly had to be a compromise
– IBM were by far the biggest and most influential
» also had a major say in what emerged…
The OSI protocol stack
Transmit
Station
AH DATA
Application
Layer
Application
Layer
DATA
AH
PH AH DATA
Presentation
Layer
Presentation
Layer
DATA
AH PH
Session
Layer
Session
Layer
DATA
AH
PH
SH
TH SH PH AH DATA
Transport
Layer
Transport
Layer
DATA
AH
PH
SH
NH
TH SH PH AH DATA
Network
Layer
Network
Layer
DATA
AH
PH
SH
NH
TH SH PH AH DATA
Data link
Layer
Data link
Layer
DATA
AH
PH
SH
Physical
Layer
Physical
Layer
SH PH AH DATA
LH
Receive
Station
LT
Link
LT
TH
TH
TH
NH
NH
LH
Focus on Levels 3 & 4
User Specifies
Service
Transport
Layer
Network
Service
Network
Layer
Network provides
Service
Network layer service definitions
Importance of levels 3 & 4

Need to be downwards compatible with the
“hardware oriented” aspects of network
messaging
– e.g. network card drivers (levels 1 & 2)

Also upwardly compatible with “user
orientated” aspects (levels 6/7), through:
– “Sockets” interface
» Originated with Unix
» Adapted for NT
– NETBIOS interface
» Preferred option for Windows applications
LAN Operating Systems
and OSI layers

Comparing levels 3,4 and 7…
– Novell Netware
» network protocol: IPX/SPX
» resources located using: NDS
– Unix (in its many forms)
» network protocol: TCP/IP
» resources located using: NFS
– Apple
» network protocol: TCP/IP
» resources located with: Open Directory
– Windows XP/2K3
» network protocol: TCP/IP
» resources located using: Active Directory
Windows Networks and Unix

Unix has been around since the mid-1970s
– been running TCP/IP almost as long
– very stable & secure

Windows NT started in 1993 (!)
– Changed to use TCP/IP
(original used a Microsoft protocol
– pitched as a direct rival for Unix, then adopted &
adapted TCP/IP making mixed environments
possible

NOW, two types peacefully co-exist
– something to do with the common protocol?
Principles of Connectivity - 1

For a client to connect to…
– another computer, the two devices need a common
protocol covering all 7 OSI layers
– a network, it does need to be able to run the default
protocol for that network

Example:
– for a Windows Server machine to connect to
a Netware server it needs to run IPX/SPX
» can easily be added to the client machine’s protocol
stack
» connection to Netware servers then possible via
hub/router and the Novell login
Principles of Connectivity - 2

Client machines on one network should also
be able to connect to resources on another
network:
– follow principle 1
– install additional software on each client machine
that will allow connection to resources using
names from the other network’s resource
database

Example: use a Netware printer from an XP
client
– set up NWLink
– set up client service for Netware (CSNW)
Principles of Connectivity - 3

Client machines on one network should
be able to connect to shared folders on
servers on another network:
– follow principle 1
– install software on the local server that will
connect to the relevant remote server(s)
– clients connect to the remote share via the
local server
Principles of Connectivity - 4

Example: use a Netware shared
directory from an XP client
– install NWLink on Windows XP clients and server
– install Gateway Services for Netware (GSNW) on
the relevant Windows 2003 Server
– the Netware share becomes available as a
Windows 2003 Server share
– local clients can link to the share use the “map“
command
Using different level 3/4
protocols in the same machine

For Netware…
– a “mixed environment” will mean IPX/SPX and
TCP/IP

Both level 3/4 protocols
– can interface directly with the same level 1/2
protocols
– each has to be connected using its own binding

Level 1/2 protocols provided by network card
software
Configuring Network Cards for
Different Level 3/4 Protocols

TCP/IP just needs IP addresses
– client IP can be allocated by DHCP from a
server
– default Internet Gateway & subnet mask should
be added manually
– Existing Internet Gateway IP address found
using ipconfig
– Link with any IP address can be checked using:
» ping <IP address>
Configuring Network Cards for
Different Level 3/4 Protocols

Netware protocols getting scarcer
– still found at UoW, at the moment…
Network cards on campus network still
configured according to “frame type”
(level 2)
 To interface with level 2 IPX/SPX needs:

– frame type (relates to IEEE standard)
– Netware Network Number (for routing)
Configuring Network Cards for
Different Level 3/4 Protocols

Using “NWLink”
– Windows equivalent of IPX/SPX
– when installing, using auto detection allows
the network card to provide the required
information by itself
» frame type on the server can be (default IEEE 802.2)
» network number
Importance of Frame Type
for IPX protocol

Level 3 protocol interfaces with a network
card at each end of the communication
– therefore essential that IPX frame type settings
are correct


Frame settings determined by the Netware
server
Can be displayed using:
– ipxroute config

NWLink should set frame type accordingly on
Windows XP machines wishing to be client to
that Netware server
Further (Deeper) aspects of
IP addressing



An IP address is made up of four fields
Each field is represented by a single byte
Divided into two sections:
– host address (first field/s)
– network address (remaining field/s

Depending on the dividing position,
categorised as different types (classes) of
address
IP Address Classes

Several address classes:
– Class A (w then x.y.z)
» For networks with many hosts
» High order bit in first field set to 0
– Class B (w.x then y.z)
» For medium sized networks
» High order bits in fields 1 and 2 set to 1-0
– Class C (w.x.y then z)
» For small LANs
» High order bits in fields 1,2,3,4 set to 1-1-0-1
Class A, B, C Addresses and first
IP field

A: First bit of first field set to zero…
– IP decimal number always <128

B: Fist bit set to 1, second to zero…
– IP decimal number always between 128 and 191

C: First and second bits set to 1, third to
zero..
– IP decimal number always between 192 and 223
Class D and E IP Addresses
Little used in practice…
 D: First, second, third bits set to 1,
fourth to zero

– IP decimal number always between 224
and 239

E: Fourth bit also set to one…
– IP address always between 240 and 255
Reserved IP addresses

127.0.0.1 to 127.255.255.255.255
– used for loopback testing

Private addresses for Internal networks:
– Class A 10.0.0.0 to 10.255.255.255
– Class B 172.16.0.0 to 172.31.255.255
– Class C 192.168.0.0 to 192.168.255.255
Subnet Masks


Purpose is to inform the hosts which bits in
the full 32-bit IP address correspond to the
network address and which bits correspond
to the host address
Value depends on whether it is a class A,
class B, or class C network:
– Class C: 255.255.255.0
– Min for Class B: 255.255.0.0
– Min for Class A: 255.0.0.0
Subnet Marks and Network Size
1subnet = 255 computers!
 255.0.0.0 subnet mask

– between 1 and 16 million

255.255.0.0 subnet mask
– between 1 and 65535

255.255.255.0 subnet mask
– between 1 and 255
Subnet Masks and Network
Packet Optimisation

Choice of an appropriate subnet mask
relates directly to efficiency of
management of network traffic
– determines packet header size
– should be a small as possible
DHCP & Packet Optimisation

Range of IP addresses set using DHCP
known as a scope
– class C network… scope would be for a maximum
of 254 addresses

Once an IP address has been assigned to a
computer, it carries a lease
– Could be leased for the duration of the session
» e.g. ISP allocation to dial up clients
– Could be leased for several days
» e.g. allocation to clients on a LAN
Thanks for listening