2 concerns:

• receiver be able to extract the same message from the signal sent by sender
• increase the efficiency of encoding - data compression

2 sides need to agree to a presentation/message format

Data Manipulation Functions

• presentation formatting
• data compression

7.1 Presentation Formatting

presentation formatting - the transforming between: application data and data suitable for transmission over a network
encoding/marshalling - application data to network data
decoding/unmarshalling - network data to application data

Issues in Argument Marshalling:

• what data types does the system support
• data conversion between different machine architectures
• identifying kinds of data within the message

What Data Types Does the System Support

data types levels (lowest to highest):

• base types - typically includes: integers, floating-point numbers, characters, ordinal types, booleans, etc
• flat types - structures and arrays
• complex types - types that are built using pointers/memory addresses

Data Conversion Between Different Machine Architectures

conversion strategies:

• canonical intermediate form
  • both sender and receiver settles on external representation for each type
  • in networking, this is known as N-by-1 solution: each N machine architectures must be able to convert between its own representation to the single external representation
• receiver-makes-right
  • sender transmit data in its own internal format; the receiver is then responsible for translating the data from sender’s format into its own format
  • in networking, this is known as N-by-N solution: each N machine architectures must be able to convert between its own representation to all N architectures

Identifying Kinds of Data Within the Message

2 common approaches:

• tagged data - is any additional information included in a message - beyond the concrete representation of the base types - that helps the receiver decode the message
• untagged data - both sender and receiver agrees on a predetermined type for each data

Stubs

• is a piece of code that implements argument marshalling
• on the client side, the stub marshals the data into a linear message
• on the server side, the stub unmarshalls the linear message into the data
• stubs can either be: interpreted or compiled

Network Data Representation Examples (XDR, ASN.1, NDR)

External Data Representation (XDR)

• is the network format used with SunRPC
• supports the entire C-type system with exception of function pointers
• defines a canonical intermediate form
• does not use tags (except to indicate array lengths)
• uses compiled stubs

Abstract Syntax Notation One (ASN.1)

• is an ISO standard
• supports C-type system without function pointers
• defines a canonical intermediate form
• uses type tags
• stubs can be either: compiled or interpreted
• used in the Internet standard Simple Network Management Protocol (SNMP)
• represents each data item with a triple of the form: ⟨tag, length, value⟩

Network Data Representation (NDR)

• data-encoding standard used in the Distributed Computing Environment (DCE)
• uses receiver-makes-right (unlike XDR and ASN.1) by inserting a tag at front of each message; individual data items are untagged
• uses compiler to generate stubs
• supports C-type system

Markup Languages

• Extensible Markup Language (XML) - takes tagged approach to the extreme
• XML Schema Document (XSD) - a schema language defined for XML
• Efficient XML Interchange (EXI) - binary XML format for greater compactness and faster parsing

XML Namespaces - solves the problem of name clashes

7.2 Multimedia Data

since multimedia data are consumed by humans we want to keep the information that is most important to a human, while getting rid of anything that doesn’t improve the human’s perception of the visual or auditory experience

compression types:

• lossless compression - guarantees data received is exactly the same as data sent
• lossy compression - does not guarantee data received is exactly the same as data sent

Lossless Compression Techniques

huffman codes

• encoding data of: higher occurrences with fewer bits and lower occurences with more bits

run length encoding (RLE)

• the idea is to replace consecutive occurrences of a given symbol with only one copy of the symbol, plus a count of how many times that symbol occur
• e.g. the string AAABBCDDDD would be encoded as 3A2B1C4D
• works well on images due to large homogeneous regions

differential pulse code modulation (DPCM)

• the idea here is to first output a reference symbol and then, for each symbol in the data, to output the difference between that symbol and the reference symbol
• e.g. using symbol A as the reference symbol, the string AAABBCDDDD would be encoded as A0001123333 because A is the same as the reference symbol, B has a difference of 1 from the reference symbol, and so on
• when the differences are small they can be encoded with fewer bits than the symbol itself. In this example, the range of differences, 0–3, can be represented with 2 bits each, rather than the 7 or 8 bits required by the full character. As soon as the difference becomes too large, a new reference symbol is selected
• works better than RLE for most digital imagery, because adjacent pixels are usually similar

delta encoding

• simply encodes symbol as the difference from the previous one
• e.g. AAABBCDDDD would be represented as A001011000
• also possible to perform RLE after delta encoding, since we might find long strings of 0s if there are many similar symbols next to each other

dictionary-based methods

• Lempel-Ziv (LZ) compression algorithm is the best known dictionary-based compression method
• the idea is to build a dictionary/table of variable-length strings that would be expected to be found in data, then replace each of these strings with the corresponding index

7.2.2 Image Representation and Compression

• images are composed of pixels
• pixels are either: grayscale or color
• there are various color spaces:
  • RGB - red, green, blue values
  • YUV - Y luminance - the brightness, U and V containing chrominance - color information

Graphical Interchange Format (GIF)

• starts with an RGB color space: 8-bits for each of the 3 color dimensions for a total of 24-bits
• then reduces the 24-bit color images to 8-bit color images: by picking the 256 colors that most closely approximate the colors used in the picture
• all 256 colors are stored in a table and each indexed with an 8-bit number
• then the value for each pixel is replaced by the appropriate index
• then runs an LZ variant over the result, treating common sequences of pixels as the strings that make up the dictionary

Joint Photographic Experts Group (JPEG)

• starts off by transforming colors from RGB space into YUV space using linear equations:
  • Y = 0.299R + 0.587G + 0.114B
  • U = (B-Y) * 0.565
  • V = (R-Y) * 0.713
• we compress U and V components more aggressively, because human eyes are less sensitive
• subsample - take a number of adjacent pixels, calculate the average U or V value and transmit that. the luminance Y component is not subsampled
• after subsampling the image, we feed it through 3 phases one 8x8 block at a time:
  • discrete cosine transform (DCT) - transforms image as spatial domain into an equivalent signal in the spatial frequency domain. this is a lossless operation
  • quantization - quantizes the resulting signal, which loses the least significant information contained in that signal
  • encoding - encodes the final result
• decompression follows these same 3 phases in reverse order

7.2.3 Video Compression

Moving Picture Experts Group (MPEG)

• each frame can be compressed using the same DCT-based technique used in JPEG
• then takes a sequence of video frames and compresses them into 3 types of frames (each frame is compressed into one of these frame types):
  • I frames (intrapicture) - can be thought of self-contained reference frames
  • P frames (predicted picture) - not self-contained, specifies the differences from the previous I frame
  • B frames (bidirectional predicted picture) - not self-contained, gives an interpolation between the previous and subsequent I or P frames

7.2.4 Transmitting MPEG over a Network

• TODO page 614

7.2.4 Audio Compression

MP3