Image Processing Reference
In-Depth Information
RS interface, which naturally only forms a small part of a complete layer definition. Essentially,
layer presents to its upper layer all that is needed to transfer a given data frame.
Layer (Data Link Layer)
This layer is a pure point-to-point connection with the task of guaranteeing transmission between
two network nodes. his firstly involves the formation of the data frame (Figure .), which typically
contains a header with control and address information and the actual data. he second task of the
data link layer is the coding and checking of the frame (e.g., via CRC: cyclic redundancy check),
to allow transmission errors to be detected or even corrected. This also includes the checking of
timeouts, or verification that the corresponding responses and confirmations are received from the
opposite side. he data link layer thus provides the following service to the above lying layer : Setup
of a logic channel to an opposite end device without intermediate nodes, or the transmission of a data
frame between two end points. In practice, layer becomes too overloaded with functions to allow
for a straightforward implementation. herefore, it is usually subdivided into the logical link control
(see IEEE .), which sets up the connection to layer (to which the error detection mechanism
is assigned) and the medium access control (MAC) to link to layer (this generally controls who is
able to transmit when).
Layer (Network Layer)
If there are nodes between the end points of an end-to-end connection, packets must be routed. In
layer , the paths between origin and destination are established via the specified target addresses.
This is easy if the corresponding path lists are available in the nodes. It becomes more complicated
when the paths are to be optimized on the basis of various criteria such as cost, quality, load, delay
times, etc., if the path conditions change during a transmission, packets need to take different paths
due to bandwidth considerations, or the packet size is unsuitable for certain paths. he task of layer
is by no means a trivial one, especially when there are various physical transmission media within
the network with different transmission speeds. It is also necessary to ensure that congestion does
notoccuralongthepaths,whichwouldcausethemaximumdelaytimestobeexceeded.
A differentiation is made between connectionless and connection-oriented services. In the case
of a connectionless service (datagram service), there is no allocation of fixed channels; every trans-
mitted package must include the complete address and is sent as an independent unit. With a
connection-oriented service (virtual circuit service), a virtual channel is made available, which
from the point of view of the user offers the advantage that the data packets need not include any
addresses. One of the first protocols of this type to be implemented and which is still in use today is
the X. (ISO ). For fieldbus systems, such virtual circuits (and therefore connection-oriented
layer- services) do not play any practical role. In any case, layer presents to its upper layer a valid
path through the network for one individual data packet.
Layer (Transport Layer)
The transport layer sets up an end-to-end connection. his means that the receiver does not route the
data further, but passes them on to layer , already prepared. here are various mechanisms available
in layer . If the data to be transmitted are too big, layer can split them up and transmit each
piece individually. If the transmission times are long or there are a number of possible transmission
paths, it is useful to number the split packets. he receiver station must then recombine the individual
packets in the right order. On layer , connection-oriented and connectionless protocols are available
as well. The most popular examples are TCP for connection-oriented and user datagram protocol
(UDP) for connectionless data transmission. hey originated in the Internet world but are becoming
increasingly important in the field-level networking domain.
