|
UNDERSTANDING THE OSI REFERENCE MODEL
The OSI reference model describes modular building blocks of various
functions that must be provided to have program to program communications
between similar or dissimilar hosts. Each functional area describes
certain tasks that must be performed by network hardware or software
in order for these network communications to take place. Not all
methods of network communication extend to all the layers of the
full OSI reference model. For example, internetwork communications
between programs on differing hosts using reliable TCP/IP communications
might use all layers of the OSI reference model, while local Windows
for Workgroups communications within the same network using the
NetBEUI protocol might have communications occurring only at the
Data Link Layer and Physical Layer levels.
One important thing to understand is that the OSI reference model
describes what must transpire for program to program communications
to occur between even dissimilar computer systems. Each layer is
responsible to provide information and pointers to the next higher
layer in the OSI Reference Model. The Application Layer (which is
the highest layer in the OSI model) makes available network services
to actual software application programs.
In describing the various functional areas of the OSI reference
model, I will start from the bottom up. However, when we describe
communications between two computer systems, we describe the communications
beginning at one computer system at the highest layer that communications
occur, going down through to the Physical Layer, through the transmission
media and network, over to the other computer at the Physical Layer,
then up through to the highest layer of communications on the other
computer system. These communications are usually symmetrical. In
the case of this type of communications occurring through a routed
internetwork, the network portion of the diagram (between the OSI
reference layers on each computer system), may also extend on through
the Physical, Datalink, and to the Network Layer.
In the diagram above, the left stack of the OSI reference model
represents computer number one, the middle two stacks indicate the
actual routed internetwork, and the rightmost stack indicates computer
number two. The arrows indicate the path taken by data sent from
a program on computer number one to a program on computer number
two, through the routed internetwork.
BACK TO TOP
The lowest layer of the OSI reference model is the Physical
Layer. It is the foundation of communications between any computer
systems. The Physical Layer specifies the transmission media, transmission
devices, network structures, and the actual data signals across
the transmission media. The transmission media provides the physical
path through which data signals flow. The transmission media consists
of either bound media (usually some type of cable) which have a
central conductor of data signals surrounded by a physical outer
jacket, or unbound media which rely on radio waves, microwaves,
and infrared signaling.
BACK TO TOP
In most LAN network communications, the Physical Layer describes
some type of cabling system as the transmission media. The Physical
Layer also describes the transmission devices that attach to the
media, the physical connector specifications, and the electrical
or optical signaling characteristics (analog or digital, signal
levels, and signal encoding methods). The Physical Layer lastly
describes the network topology, and how the Physical Layer transmission
media is to be distributed (i.e., in a bus, star, ring, mesh, etc.,
topology).
BACK TO TOP
The most simple type of communications between computer systems
(serial or parallel) occurs at the Physical Layer, where the Physical
Layer would consist of the transmission media, transmission devices,
a simple point-to-point network structure, and network signaling
and interface specifications (such as an RS-232 serial communications).
BACK TO TOP
The second layer in the OSI reference model is the Datalink
Layer. This layer describes the functions that must occur for
LAN communications within a local network. The Datalink Layer provides
the first level of organization of the data bits into a rudimentary
structure called a Datalink frame. This frame is organized into
fields of information that convey the beginning and ending of the
frame, the address of the sender, the address of the receiver, a
method to ensure that the frame does not contain errors that may
have occurred in the course of being transmitted through the transmission
media, and an area to provide some basic administrative function
(such as flow control, frame length calculations, and protocol decisions).
Datalink Layer flow control manages how much data the destination
system can handle.
BACK TO TOP
The Datalink Layer is broken into two logical areas, the Media
Access Control (MAC) sublayer, and the Logical Link Control
(LLC) Sublayer. The MAC sublayer refers to the Media Access
Protocol (how do stations on a network gain access to the media
and permission to transmit their data - contention, token passing,
polling), and the physical addressing of stations on the network.
This would include the Source and Destination address sections of
the Datalink frame. The LLC sublayer includes portions of the Datalink
frame administrative or control fields and CRC fields that are responsible
for the frame synchronization, flow control, and error checking
within the frame.
BACK TO TOP
BACK TO TOP
The physical addressing at the Datalink layer (used for source
and destination addressing in a Datalink Layer frame) is called
a physical address because this address (also called a
MAC-layer address) is hard coded into the network interface in their
computer or workstation. This addressing is useful for conveying
all the information necessary to direct information to and from
stations within a local network. The addressing is usually assigned
by the manufacturer of the network interface at the time of manufacture,
and the first half of the address contains hex address information
that is unique to each manufacturer. The last half of the MAC layer
address is unique to that individual NIC card. Networks with duplicate
addresses can have problem s in knowing where to properly deliver
their data. This is why all manufacturers purchase unique identifiers
and blocks of unique addresses for all the network interfaces that
they manufacture.
BACK TO TOP
When we refer to Ethernet, or Token Ring, or other LAN Networking
systems (ARCNET, LocalTalk, FDDI, etc.), we are referring to the
Physical and Datalink layer descriptions of these networks. For
example, an Ethernet workstation will construct datalink layer frames
that may conform to the IEEE 802.3 specification for Ethernet frames,
and which follow that specifications requirements for signaling
types, levels, media, physical connectors, etc. The Physical and
Datalink Layers also specify the Media Access Control methodology
(CSMA/CD), Ethernet topologies, and other rules necessary for ensuring
a functional network design.
BACK TO TOP
The Network Layer is responsible for the internetworking process
that needs to occur to reliably send and receive data between networks.
It provides logical addressing at the Network Layer (as this addressing
is usually specified in software, and is not hard coded into network
interfaces as it is at the Datalink Layer), and provides for network
routing, flow control, sequencing, and translation functions. Network
Layer flow control monitors network connection. Network layer addressing
provides addressing that is specific to the logical addressing assigned
to a particular network protocol. For example, the Ethernet workstation
described in the previous paragraph may have Ethernet as the Physical
and Datalink layer network type, and may run both TCP/IP and Novell
IPX/SPX as different network layer protocols. Each network layer protocol
will have a different logical addressing scheme to enable communication
between networks.
BACK TO TOP
This brings up an interesting issue involving network protocols.
Without going too deep into the subject, there are some common protocols
used in LAN networking that do not extend to the network layer.
These include DEC LAT, NetBIOS, NetBEUI, and others. Because they
have no provisions for network layer addressing and routing, connecting
these protocols between different LAN's must be done with MAC layer
devices such as LAN bridges or LAN switches. Protocols that operate
at the network layer (such as TCP/IP, IPX/SPX, DECNET, etc.) utilize
routers to interconnect networks. Most routers have the ability
to route routeable protocols and to bridge non-routeable traffic.
BACK TO TOP
The Transport Layer provides reliable end-to-end communications
by providing service addressing, flow control, datagram segmentation,
and end-to-end error checking. The Transport layer is the second
of three error checking levels in the OSI model (error checking
occurs at the DataLink Layer, Transport, and Session Layer). While
the Network Layer is responsible for transporting data from Network
A to Network B, the transport layer ensures that packets arrives
(or is reassembled) in one piece, and make sure that the data is
directed to the appropriate service. While I have explained physical
MAC-Layer addressing which occurs at the Datalink Layer, and the
logical addressing at the network layer, we will now see that the
Transport Layer concerns it self with the next layer of addressing
described in the OSI model, which is Service Addressing. Service
addressing identifies addresses or ports which point to upper layer
network services. The transport level addressing also keeps track
of multiple connections or conversations which might occur on a
network attached computer system. It does this by tracking a connection
identifier (connection ID, port, or socket), or by a transaction
identifier (which would track each request, as opposed to tracking
a conversation). The Transport Layer is also responsible for breaking
larger units of data sent down from higher layers into smaller pieces
that can be transported across the network. These pieces would then
be reassembled at the Transport Layer of the receiving computer
system and passed on to that computers higher layers. This is accomplished
by assigning segment numbers to fragments of a common conversation.
The Transport Layer accomplishes it's contribution to the flow control
process by negotiating window sizes.
BACK TO TOP
The Session Layer is responsible for connection establishment,
data transfer, and for connection release. In this way the Session
Layer provides mechanisms which establish, maintain, synchronize,
and manage communications between computer systems. The Session
layer is responsible for the establishment of connection ID numbers,
and relies on the transport layer to provide the information which
identifies the correct services. The Session Layer is responsible
for the coordination of acknowledgement numbering and retransmission
procedures. It tracks who initially initiated a conversation. Session
layer is responsible for re-establishing a logical communications
session between computers should that session be prematurely terminated
(due to failure at a lower layer in the communications process).
The Session layer would either resume the interrupted dialog, or
initiates another new session to the other computer system. The
Session Layer also manages planned connection release of a communications
session.
BACK TO TOP
The Presentation Layer is responsible for translating data
into formats that can be readily understood by each computer system.
The Presentation Layer concerns itself with translating differences
at the bit level, byte level, or character level, and with file
syntax. An example of this might be translating data formats between
two different computer systems using the TCP/IP protocol, that have
different data conventions. One host might use ASCII code and may
have the left-most bit be the highest order bit (or most significant
bit) in a byte of data, while the other host system utilizes EBCDIC
and for which the left most bit is the lowest order bit (least significant
bit). Computers need to agree the method of identifying how many
bits equals a whole character and what file syntax is used by each
system (file organization, beginning and ending boundaries, naming
conventions, read and write access security, and file storage methods).
The Presentation Layer also concerns itself with data compression
and encryption.
BACK TO TOP
Probably the most misunderstood layer of the OSI model is the Application
Layer. Contrary to what many people think, this IS NOT
the layer where the software application (such as your spreadsheet
or word processing program) resides. It is the layer where network
applications that facilitate network connectivity reside (file transfer/FTP,
virtual terminal/Telnet, and electronic mail, for example), and
where service advertisement and service availability are managed.
Service Advertisement is the process of letting systems know which
services are available on the network. Once it has been made advertised,
it must be made available to the computers local operating system.
BACK TO TOP
Network services include file services, print services, message
services, application services, and database services. Network layer
services are provided at the Application Layer. The Application
Layer makes network services available to a computers local operating
system. This is accomplished through Operating System call interception,
remote operation, or via collaborative availability. In OS call
interception, the local operating system is completely unaware of
the existence of network services. In remote operation the local
system is aware of the network, but the network server is unaware
of the client. In collaborative systems, both the server and the
client are aware of each other and work together to coordinate service
availability. This method is typically used in peer-to-peer network
operating systems.
© MSI Communications 1998
BACK TO TOP
|
