Microtronix Access User Guide
11 Conversion and Encapsulations
When converting from X.25 to TCP, a fundamental difference in the way application data is transferred
needs to be taken into consideration and accounted for. X.25 is message oriented, and TCP is byte-
stream oriented. In order to preserve message boundaries required by X.25 applications, a TCP
application must introduce a message boundary preservation method into the TCP byte stream so that
the intervening X.25/TCP gateway is able to convert the contiguous stream of messages back into
discreet messages on the X.25 side. The TCP application and the X.25/TCP gateway must agree on a
common encapsulation method. Without this agreement, the X.25/TCP Gateway cannot distinguish
message boundaries in the TCP data.
Encapsulation of application messages is often required so that a receiver application can reconstruct a
message based on the boundaries defined by a transmitting application. Encapsulation is necessary
because often the application message is larger than the network's maximum transmitted packet size,
and is fragmented across multiple packets, or multiple messages merged within a packet. This means
that the network packet is an unreliable boundary for application messages.
With X.25 interfaces, the boundary offered by the X.25 protocol is the use of a MORE data bit in the
data packets to mark when there is continuation of a message in the next packet. The last part of the
message is in a data packet with the MORE bit set to zero. With MORE bit markers, X.25
implementations can receive an entire, discreet message in a single read, and transmit a message in a
single write.
TCP, on the other hand does not have a specific boundary around messages sent by applications. Like
an asynchronous serial data stream, TCP is byte-stream oriented. Even if an application sends a
message as a single write operation, there is no guarantee that the message arrives at the receiver
application in a single discreet segment. It can get split across TCP/IP packets, or merged with
preceding or following messages. So TCP/IP applications must add their own message boundary
markers if the message is to be preserved. This method needs to be supported by the X.25/TCP
gateway in order for it to be able to translate or convert the message from that encapsulation method to
the M-bit method towards the X.25 application.
11.1 RAW
Not all applications require the boundaries of the message to be preserved by an intervening X.25/TCP
gateway. The content of the message provides the appropriate markers like start and/or end
characters, or has a message header that already contains a length indication. This only works for
X.25 applications that do not rely on the message being contained in a MORE-bit sequence of packets.
With RAW conversion, both the X.25 and the TCP data are treated as byte streams, and the Gateway
does not add MORE bits to data packets. It makes no effort to preserve X.25 message boundaries.
This method requires that both the X.25 application and the TCP/IP application support the embedded
boundaries of the application message, or simply don't require any boundaries.
11.2 RAW-MBIT
This method of conversion in the X.25/TCP gateway makes a “best-effort” to preserve boundaries of
messages without introducing any additional data into the TCP byte stream.
Requirements:
●
The X.25 application uses More-bit packet sequences to mark message boundaries.
77
