242 Practical TCP/IP and Ethernet Networking
15.4.2 TCP/IP based factory automation
Schneider Automation:
Richard Hirschmann Gmbh:
Allen-Bradley:
15.4.3 Thin servers
LANTronix,
15.4.4 Web compatible SCADA systems
WizNet,
15.4.5 Java
RCS-7 Java,

16
/TYZGRROTMGTJZXU[HRKYNUUZOTM
+ZNKXTKZY_YZKSY
 5HPKIZO\KY
When you have completed study of this chapter you should be able to:
• Define the functions of various types of network driver software
• Describe the parameters which need to be set for a network card to
function correctly
• Determine how network cards are configured under the plug and play and
PCMCIA architectures
• Specify the uses for a protocol analyzer
 4KZ]UXQJXO\KXY
 4KZ]UXQJXO\KXY
The network driver is a program used to provide an interface between the network card
and the higher-level protocols. It spans the data link and network layers of the OSI model
as shown in Figure 16.1.
The network driver needs to match the specific hardware configuration of the network
card such as its addresses of I/O ports, control and status registers, etc. Ethernet cards use
the same IEEE 802.3 protocol so they can communicate with one another but each needs

a unique driver because of the different vendor hardware implementations.
 )USVGZOHOROZ_GTJ[YGMK
The network driver must also be compatible with the appropriate network operating
system protocols in the network, transport and session layers that are used to send data
6XGIZOIGR:)6/6GTJ+ZNKXTKZ4KZ]UXQOTM
across the network. In early systems changing from one network protocol to another, e.g.
TCP/IP to SPX/IPX, generally necessitated changing the network driver. The network
operating systems communication protocols and their relationship to the OSI model are
shown in Figure 16.1.


Figure 16.1
Network operating system drivers/protocols
 4KZ]UXQIGXJJXO\KXIUTLOM[XGZOUT
The configuration of the network card must be set at installation to avoid conflict with the
other devices installed on your computer. The manufacturers usually provide an
installation guide and/or configuration software to help you set the correct options. The
main parameters that may need to be set are as follows:
/87INGTTKR
The IRQ channel sets a unique interrupt request (IRQ) vector for when the network card
needs attention. This must not conflict with existing devices. It needs to be checked in
the normal operating environment. For example, with Windows applications the network
card configuration software should be run from within the Windows shell.
*3'YNGXKJSKSUX_
The Network card communicates its data to computer using either direct memory access
(DMA) or use of a block of shared memory. The base address of the shared memory
(usually 64 kilobytes) is defined here. Network cards are designed to be flexibly
reconfigurable to accommodate other applications on your computer.
8'3HGYKGJJXKYY
RAM base address defines the beginning of the address space (usually 16 kilobytes) used

by the network card. Other devices must not use such address space. For example,
extended memory manager software should be set to exclude such memory to avoid
conflicts.
/5HGYKGJJXKYY
I/O base address defines the beginning of the address space used to communicate with the
internal registers on the network card. Avoid conflicts with existing devices.
/TYZGRROTMGTJZXU[HRKYNUUZOTM+ZNKXTKZY_YZKSY
853HGYKGJJXKYYGTJYO`K
This is used for systems with an auto-boot ROM to configure diskless workstations.
These load their operating systems over the network. Match the base address and
EPROM size to the supplied boot ROM.
The network card configuration is set on the card using switches or links and/or stored
in flash memory (EEPROM) on the card by the configuration software.
 4KZ]UXQJXO\KXOTZKXLGIKYVKIOLOIGZOUT4*/9
This is a software specification developed by 3COM and Microsoft in 1988 for use
mainly in DOS and OS/2 operating systems. This defines a standard interface for
communication between the MAC layer and any compatible protocols. This means that
the MAC driver in any vendor’s NDIS compatible network cards can pass data to any
NDIS compatible protocol. This standardization enables products from various vendors to
be interconnected and to provide simultaneous support for multiple protocols. These
NDIS drivers can be unloaded from memory to conserve DOS RAM space or allow
changes to other drivers.
 5VKTJGZGROTQOTZKXLGIK5*/
The ODI architecture provides an alternative to the OSI layering structure that can allow
a number of network cards to simultaneously support different protocol stacks such as
TCP/IP, SPX/IPX, etc.
The ODI architecture makes use of a link support layer (LSL) and a multiple link
interface driver (MLID) as shown in Figure 16.2. The MLID corresponds to part of the
data link layer and interfaces to the LSL, which covers part of both the data link and
network layers. This provides a standard, hardware independent, virtual interface for the

network cards. The LSL switches multiple protocol packets to the correct MLID or the
correct protocol stack as required.


Figure 16.2
ODI architecture
6XGIZOIGR:)6/6GTJ+ZNKXTKZ4KZ]UXQOTM
 6GIQKZJXO\KXY
The packet driver is the generic interface between the TCP/IP protocol stack and the
software responsible for the local area network card hardware. The packet driver hides
the hardware specifics from the protocol stacks and likewise hides the protocol issues
from the hardware. The packet driver operates at the MAC layer, and its implementation
is critically dependent on the specific card hardware. Common driver specifications have
evolved to provide standard interfaces to both the protocol stack and the card hardware,
thus enabling the protocol to run on top of all network cards, which have that MAC driver
specification.
A common example is the packet driver developed by FTP Software, Inc, in 1987. The
specification defines how the MAC driver loads and operates under DOS and defines a
common software interface for various protocol stacks. The network card is independent
of the protocol stacks, and one card can simultaneously handle packets destined for
multiple protocols. Each protocol uses a software interrupt in the range 60 h–80 h to
communicate with the packet driver.
 6R[MGTJVRG_GXINOZKIZ[XK
Modern PC motherboards support the PC/ISA plug and play (PnP) architecture. With this
architecture the PC identifies which slot the particular card is inserted into and the BIOS
dynamically assigns the resources the card requires e.g. IRQ, DMA channel etc, resolving
any resource conflicts. These resource details are registered in the ESCD (extended
system configuration data) and stored in flash memory on the motherboard. The data
structure defines the resources used by each device and card on the system. When using
PnP it is important to register all legacy cards and devices (non PnP), otherwise resource

conflicts are likely.
 6)3)/'OTZKXLGIK
 /TZXUJ[IZOUT
PCMCIA are the initials of the Personal Computer Memory Card International
Association, which was formed in 1989 to promote the standardization and
interchangeability of PC cards. As the name indicates initial devices were memory cards
implemented as ‘virtual disk drives’ for mobile computer support. PC cards now come in
a wide range of memory devices such as RAM, ROM FLASH memory, AT Attachment
(ATA) hard drives, and many I/O devices including modems and network interface cards.
The interface enables these devices to be powered from the computer and automatically
detected by the system as soon as they are installed and then automatically configured.
This gives the PC cards the ability to be inserted into a PCMCIA socket after the system
has already been powered up.
 6)3)/'OTZKXLGIK
The PCMCIA interface consists of the following, as shown in Figure 16.3:
.GXJ]GXK
• The 16-bit PC card, (A 32-bit PC card and socket interface also exists
called CardBus)
• PCMCIA socket