Sending and receiving data
There are several ways to transfer data. No matter if it is a file, an image, textual or raw data, it is all the same. What is needed is a common language between the data sender and the data receiver, in order for each to understand the other. This language defines a format. Formatting data is a way to "shape" it, describing its beginning and its end, in order for the receiver to know where data starts and ends.
Many formats are of course possible, mainly depending upon the kind of data itself. Basically, we can consider that we have two kinds of data :
- predictable data content
- unpredictable data content
In the first case, we know what kind of content the data contains - ie the byte set that can be used is known (byte values from 32 to 127, for example). In the other case, the data content is not known - ie the byte set can be any possible byte value (from 0 to 255). This major difference will mostly determine the format of the data to be exchanged, defining then the protocol.
Concerning Predictable Data Content, as the content values are known by advance - a subset of the 256 possible byte values, it is possible to know which byte values cannot be part of the data. This hint let us the possibility to create a delimiter made of one or more of these "extra data" byte values. This delimiter, sent after data, may then help us to know where data ends. This mode where a delimiter is placed at the end of data is named LineMode.
Concerning Unpredictable Data Content, as the content values are never known by advance - byte set values starts at 0 and ends at 255, if we plan to send several data packet (for example several records) at once, then we have to tell the receiver the size of the data we're going to send for each packet. This mode to "format" data packet by prefixing it with its size is call PacketMode.
And if we plan to send at the same time some Predictable Data Content and some Unpredictable Data Content, we can then imagine a combination of both mode.
Line mode is the most widely used, specially with ascii based protocols. But there are many protocols where line mode is used with unpredictable data. In that case that data has to converted into a specific range of bytes. The most used encoding mechanism are described here.
Base64 is an encoding process using a 64 letter alphabet where each letter representing 6 bits in the input stream. It is described in RFC 2045. All encoding and decoding procedures can be found in unit MimeUtil. The encoded data is about 33 percent larger and not human readable. Every character not used in the Base64 alphabet can be used as control character, including line end.
ASCII-hex is used in many protocols. Every character is converted into his hexadecimal equivalent and sent as such. For example the string '123' is sent as '313233'. The encoded data is twice as long and difficult human readable. Every character except 0..9, A..F can be used as control character, including line end.
Escaping is very often used. Control characters including line end has to be chosen in a way they are as less as possible in the original data. The principle is to precede the control characters with an escape character and replace them by other characters. Very often a NULL character is escaped as well. Most of the time the encoded data is only a little longer than original and good human readable.
The Quoted-Printable encoding as specified in RFC 1521 is intended to represent data which is largely human readable. It encodes certain byte ranges into their hexadecimal presentation. The encoded data remains good human readable. Procedures to encode and decode Quoted-Printable can be found in unit MimeUtil. Every encoded character can be used as control character, including line end.
Packet mode is also very widely used and most of the time designed for a specific protocol. In packet mode the data has a specific structure which eventually can be different depending on the type of packet even within the same protocol. The most universal ones are explained here.
A very often used technique is to precede the data with a header. Presides other control information this header has a field containing the length of the data.
Less used but also a very good technique is to represent the length field in hexadecimal format of 2, 4 or 8 bytes. Advantage is that it is human readable.
Fixed length data
Fixed length data is where all packets of the same type have the same structure and length. Mostly the type of the structure is indicated in a field somewhere at the beginning. If such a structure contains a data field of variable length then eater the data is padded or its length is indicated in a field in the structure.
- A want to send data, so he tell it to B during negotiation, with ascii commands using LineMode, including the size.
- After receiving that information, B switch LineMode off then tell it to A.
- When A receive this confirmation it start sending exact what has negotiated.
- When B has received it all, then he switch LineMode back on then tell it to A.
- After receiving that information both are again in negotiation phase.
Difficult to explain something :)
Wilfried 20:18, 21 February 2006 (CET)