D2D: Digital Archive to MPEG21 DIDL

1
D2D Digital Archive to MPEG-21 DIDL

• Suchitra Manepalli

2
Overview

• Digital Preservation
• Analysis Techniques
• MPEG-21 DIDL
• D2D Tool
• Qualitative Measures
• Demo

3
Digital Item

• An atomic unit of information in digital form
• a document
• an image
• a video
• Item may refer external entities
• not part of the digital item
• Current context
• preservation intent
• accessible for analysis
• digital archive

4
Digital Archive

• Collection of digital items
• collective meaning may exist
• Archive exists
• a discrete set of items
• directory of a file system
• cohesively with help of external programs
• tar, gunzip etc

5
Digital Preservation

• Preserving the digital items
• byte stream
• metadata
• reproduce the original digital item
• Continued access of digital item contents
  invariant of time and technology

6
Metadata

• Perform various analysis
• to gather information
• to verify the integrity
• to understand clearly
• Prepare a container
• to store digital item
• to store metadata

7
PDI

• Preservation Description Information
• Provenance
• Context
• Reference
• Fixity

8
Provenance

• History of a digital item
• When
• Who
• Maintain the integrity of the digital items past

9
Context

• Scope of a digital item
• Environment
• How

10
Reference

• Identification systems used for digital item
• Typical reference includes assigned identifiers

11
Fixity

• Maintains the integrity of the digital content
• Refrains modification of content information from
  unauthorized alterations

12
Example - PDI
13
Analyzation

• Applying various analyzation techniques on the
  digital items
• produces metadata
• Metadata - primary source for understanding the
  content

14
Archive Content Analysis

• JHOVE
• FRED
• File
• Strings
• Checksum

15
JHOVE

• JHOVE is JSTOR/Harvard Object Validation
  Environment
• Format specific identification
• format of the digital item
• Format validation
• conformance to standards of the format
• Format characterization
• significant characteristics of the digital item

16
JHOVE Output
17
JHOVE PDI
18
File

• Digital item classified
• purpose content storage block
• type machine-language file
• format file format or the language
• file command on UNIX machines
• file system tests
• magic number tests
• language tests

19
File Output
20
File PDI
21
Raw Characters

• Digital item consists
• bit stream
• Character set or the encoding
• ASCII (7-bit), UTF-8, UTF-16, octet-stream
• strings command on UNIX
• printable characters from the specified file,
• regardless of the file format

22
Strings Output
23
Strings PDI
24
Checksums

• MD5
• a cryptographic hash algorithm
• produces a 128-bit hash value
• idempotency of the message digest
• integrity of the original data object content
• UNIX command line utility
• md5sum is used to compute

25
Checksums Output
26
Checksums PDI
27
FRED

• Format Registry Demonstration
• demonstrates a simple format registry service
• proof of concept prototype for GDFR
• brother of TOM
• Provides information regarding registry formats

28
FRED Output
29
FRED PDI
30
MPEG-21

• Open standards-based multimedia framework
• Support interoperability of content across
  communities
• Managed, described, exchanged, collected
• Content Metadata Digital Item
• MPEG-21 Digital Item Declaration
• data model describe the set of abstract terms, or
  concepts to define digital item

31
DID Data Model

• Container
• Item
• Component
• Anchor
• Descriptor
• Condition
• Choice

• Selection
• Annotation
• Assertion
• Resource
• Fragment
• Statement
• Predicate

32
Abstract Entities

• Container
• Multiple containers or items or combination of
  both
• Item
• Encapsulates the items or containers
• Component
• Encapsulates Resources
• Specify control information for resources defined

33
Abstract Entities

• Descriptor
• Express relationship between entities
• Define the specific entity
• Resource
• Star of DID
• Digital Object or a Physical Object
• Statement
• Not part of resource
• Control, descriptive, identification information

34
DID Data Model
35
DAAP Data Model

• Signature
• specifies the type of analyzer
• environment
• versioning information
• Provenance
• type and value associated with each digital
  items history
• Context
• metadata about digital items existence
• Reference
• digital items identity

36
DAAP Data Model

• Fixity
• information that remains invariant to time.
• Raw output
• capture the analysis from the techniques as is
• Type
• elements defined for provenance, context,
  reference and fixity.
• Value
• associated with the corresponding type.

37
DIDL View
38
D2D Tool - Overview

• Implemented in Java
• Command line interface
• Tool performs functions in three discrete realms
• Submission
• Analysis
• Dissemination

39
Tool Usage

• Usage
• D2D ltarchivegt ltarchive formatgt ltcomma separated
  analyzersgt ltdidl namegt
• Supported Analyzers
• Jhove Strings FRED
• File checksum
• Archive Formats
• tar tar.gz
• gz tar.bz
• bz zip

40
Design
41
API Structure - Submission

• Identify submitted archive format
• supports a few archiving formats namely tar, zip,
  gunzip, bunzip etc.
• Explode the archive
• Initializes Analysis Manager
• driven by various analysis techniques passed as
  a parameter
• Initialize the digital item objects
• run-time identity
• referencing the file system location
• identifying the file size
• assigning the identifier

42
API Structure Analysis

• Analysis manager performs the requested
  techniques
• Presently supports JHOVE, FRED, File, Checksum,
  Strings
• Easy addition of new analysis technique
• the manager automatically applies
• Responsible for producing the timing report
• individual analyzers
• overall analysis
• Serialize the analyzed digital item
• creation of MPEG-21 DIDL components
• synching them to disk

43
API - Dissemination

• Serialized digital items are read
• Concatenated to produce the complete DIDL
• DIDL produced is well-formed and valid

44
Performance of D2D Tool

• Individual analysis timing
• Aggregate timing

45
Performance of D2D Tool

• Individual analysis timing
• Aggregate timing

46
JHOVE
encountered the XML files
47
File
48
Raw Characters
49
Performance of D2D Tool

• Individual analysis timing
• Aggregate timing

50
Archive Size
51
Number of Files
52
Performance

• Number of files Time
• Archive size Time

53
Demo
54
Conclusion

• Analysis on the digital archive
• valuable for digital preservation
• D2D tool
• proof-of-concept for other future implementations
• Timing analysis performed
• understand the amount of computational resources

55
Future Work

• Other facets of performance should be explored
• Preserve archive
• Existing XML APIs are slower for processing XML
  data
• Tool should be extended to a multiple-node
  environment like JINI

56
Questions Comments