This chapter consists of the following sections:
Multimedia in museums will rely on data interchange and telecommunications protocol and standards. In 1993, CIMI (Computer Interchange of Museum Information) concluded that underlying data representation standards such as those quoted below should be followed by museums regardless of the interchange application.
These main underlying standards are:
ASCII for text, CCITT Group 3 and 4 facsimile, JPEG (Joint Photographic Expert Group), CGM (Computer Graphics Metafile), TIFF (Tagged Image File Format), BMP (Windows BitMaP) for graphics and still images, MPEG (Moving Picture Expert Group) for video, CD Audio for audio, MHEG (Multimedia Hypermedia Expert Group) for composite object.
CIMI recognizes, however, that while the framework may seem definitive on this issue, concrete recommendations are often difficult to formulate because standards, especially in the area of video compression, are undergoing rapid development and implementations are lagging behind the adoption of new standards.
CIMI identified three approaches to interchange of database records for construction of shared catalogues:
- ISO 2709 (Information Interchange) - ISO 8824 (ASN.1 Š Abstract Syntax Notation One) and its Basic Encoding Rules (ISO 8825) -ISO 8879 (SGML Standard Generalized MarkŠup Language).
Standard Definition Description ISO 8824 &ISO 8825 is flexible technical language for specifying data structures. Together with its Basic Encoding Rules, it provides a means of specifying, encoding and decoding messages to be transmitted. The semantics are very general (Boolean, string, number ). Mappings are defined by the user.
ISO 2709 is representative of bibliographic and information (textual, descriptive data) interchange formats. It is specifically intended for communication between data processing systems. It is a content independent specification of fielded (tagged) text for data exchange.
A record contains both data for transmission and information about the data, its structure and organization for the data processing system to use. CSGML/ ISO 8879 is designed to describe how text should be formatted, printed, processed or represented in a system independent way (formatting and presentation capabilities are very basic [see Data presentation]). With its associated standard Hytime (Hypermedia/TIME-based document structuring language ISO 10744) are methods to denote the structure of free, tagged, or formatted text and to link these with associated images and sounds.
ISO 8879 is defined two parts : a Prologue and a Document instance. The prologue contains the SGML declaration (specifies facts about character set, delimiter codes and length of identifiers) and a document type declaration (DTD description of the entities of the document). The document instance is the content.
After analysis, CIMI concluded that each set of standards had strengths and weaknesses:
ASN.1 is the preferred OSI language for representing other protocols and standards. For example, ISO 8879 SGML uses ASN.1 for encoding the formats. While ASN.1 can be used to encode data that is formatted for interchange according to other standards such as SGML/SDIF, it is also possible to use it directly for both formatting and encoding.
Use of ISO 2709 has declined over the last few years especially because the adoption of MARC (Machine Readable Cataloguing) for bibliographic use. Numerous bibliographic but also non-bibliographic MARC formats are in use in over 20 countries and a bridge (USMARC) has been implemented between them. The scope of bibliographic items includes books, computer files, maps, music, visual materials.
Use of SGML is interesting for most object descriptions, particularly if it involves use of extended text in fields in which the contents of that text should be identified or if it involves text associated with graphics, images, sounds or multimedia, or if the ultimate purpose of the interchange involves paper or electronic publishing.
The suite of technical standards for Search and Retrieval which were developed as NISO Z39.50 and have since been adopted as ISO 10162/10163, are considered the most widely accepted method for searching remote systems running diverse software.
It is important to distinguish between information retrieval functions of ISO 10162/63 and the database management functions of Structured Query Language (SQL) and Remote Database Access (RDA) extensions to SQL specified as ISO 9579 RDA. ISO 10162/63 is preferred for information retrieval services where knowledge, or data in context is wanted. For example ISO 10162/63 allows searches for concepts like "Author" or "Title" to be conducted independently of the way data is represented internally in the application, or whether it is stored as a flat file or relational table. In contrast to that, SQL and RDA extensions are designed for performing database management services such as updating, archiving, or moving databases, for query of relational tables, or for distributed database applications.
Beyond the data structure definition, formatting and data presentation have to be defined. For instance, formatting and presentation are not handled by SGML in a single context (they are in fact, but with very basic capabilities). So additional standards have been defined to support the control of the way documents are represented to the recipients in document interchange : DSSSL (Document Style Semantics and Specification Language ISO 10179) for formatting documents and SPDL (Standardized Page Description Language) for page description, for instance. But they have yet to be broadly implemented in commercial software.
Nonetheless, tools are available to process SGML documents into electronic or printed output complete with user-definable formatting. Integrated tools allow changes to structure, content, or formatting at any point in the production process. The SGML set of standards seems to be the more widely accepted. Nonetheless, another standard exists in parallel that is more dedicated to office application : ODA (Office Document Architecture ISO 8613).
The basic strategies for protecting communicated information are the physical security, the authentification and the encryption. Physical security concentrates on the control of access to networks, computers or sectors of storage media such as hard disks. This security is not the subject of this chapter. Readers are referred to the literature on LAN's, WAN's and MAN's.
ANSI defined standards for authentification such as ANSI X9.9 and X9.17. They are mainly used in banking environment.
Algorithms have been designed for encryption : the two most widely recognized are DES (Data Encryption Standard) and RSA (Rivest, Shamir and Adelman).
Billing for access to data and statistics of actual use are closely linked : billing can be carried out from statistics. The minimum necessary data to calculate the statistics are:
- the identification of the incoming log (transaction ID, type of user) - the current date and time - the starting time and the ending time of the connection - the identification of the query (number of results, type of the content: text, image, video, sound; size of the data transferred)
Data on all users are preferably stored in monthly statistical files, for which privacy is ensured.
OSI Model and Definition
OSI or Open Systems Interconnection is a model defined by ISO (International Standard Organization). Its aim is to enable different computers from different vendors to communicate and exchange data transparently. The model relates to Local Area Networks (LAN), Metropolitan Area Networks (MAN) and Wide Area Networks (WAN).
It includes 7 layers or more exactly 6 protocol layers and a physical interface as shown in the figure here below.
Layer 7 : Application Data processing (file transfer, consultation)
Layer 6 : Presentation Data representation (coding,syntax, etc)
Layer 5 : Session Manages and synchronises the dialogues (e.g. password verification)
Layer 4 : Transport Assurance of reliable and transparent data transport communication
Layer 3 : Network Routing through switching nodes
Layer 2 : Data link Data transfer including correction of binary errors
Layer 1 : Physical Definition of the physical interface between network and terminal.
The OSI model is a full set of standards each concerning one of (or part of) the layers. In the following paragraphs the most important standards for respectively LAN, MAN and WAN will be discussed.
LANs (Local Area Network)
Ethernet or CSMA-CD (Carrier Sense Multiple Access Collision Detection): Ethernet is a standard for LAN. The specification is laid down in IEEE 802.3 or ISO 8802.3. It corresponds to layer 1 and part of layer 2 of the OSI model.
Layer 2 : Data link LLC (LogicalLink Control) MAC (Medium Access Control) : CSMA-CD Ethernet
Layer 1 : Physical Cable in bus Ethernet
The physical medium most widely in use is either coaxial cable (10 base 5, 10 base 2), or unshielded twisted pair (10 base T) or optical fibre organised in bus.
The theoretical transfer rate is 10 Mbits/s but the actual rate is approximately 3 Mbits/s. This is due to the protocol CSMA-CD. Before sending a signal or bit stream, a station listens to the network and only sends if the network is free. However, sometimes collisions take place when two or more stations send at the same time. These stations then have to re-send, which takes time. Henceforth the low actual transfer rate.
Enhancements are being realized in order to make Ethernet deterministic and to assure a higher actual transfer rate to the effect that Ethernet may be used for real time applications. In particular the deterministic algorithm DCR (Deterministic Collision Resolution) offers approximately 8 Mbits/s actual transfer rate.
In addition, an Ethernet 100Mbits/s standard is currently under development. Hewlett Packard and Tricom are the main players in this new development.
Token ring
Token ring is also a standard for LAN. The specification is laid down in IEEE 802.5 or ISO 8802.5. It corresponds to layer 1 and part of layer 2 of the OSI model.
Layer 2 : Data link LLC (Logical Link Control) MAC (Medium Access Control) : Token Token ring
Layer 1 : Physical Cable in ring
The physical medium currently in use is either coaxial cable (10 base 5, 10 base 2), or unshielded twisted pair (10 base T) or optical fibre organized in ring.
The theoretical rate is 16 Mbits/s. Access is deterministic : a token runs from a station to another on the ring and transmits the data. The RAMA project has chosen token ring for video play.
MAN (Metropolitan Area Network)
FDDI (Fibre Distributed Data Interface)
FDDI is standardized as ISO 9314. It corresponds to layer 1 and part of the layer 2 of the OSI model.
Layer 2 : Data link LLC (Logical Link Control) MAC (Medium Access Control) : Timed Token FDDI
Layer 1 : Physical Cable in double ring FDDI
FDDI was originally designed to work with optical fibre. Therefore it was insensitive to electro-magnetic interferences. The main drawback was the high price of optical fibre and the necessary adapter. However, it also can work with STP (Shielded Twisted Pair) or UTP (Unshielded Twisted Pair). The theoretical transfer rate is 100 MBits/s. Access resembles token ring with the principle of the timed token. The topology in a double ring offers a better security however. In practice FDDI is mainly used to link LAN's (Ethernet, Token ring). Thus it is known as a high speed backbone network.
WAN (Wide Area Network)
Packet Switching Networks CCITT X25
The X25 protocol refers only to the layers 1 to 3 of the OSI model. It is mainly used within the framework of a WAN (Wide Area Network).
Layer 3 : Network Establishment of connection, management of data transfer and release of connection
Layer 2 : Data link H.D.L.C protocol (High Level Data Link Control) assures data transfer without binary errors. It corrects errors and controls the data flow
Layer 1 : PhysicalThe interface is defined by X21 protocol and the medium is usually a leased telephone line (public network)
ISDN (Integrated Services Digital Network)
ISDN definition is based on 3 main pillars: digital connectivity from end to end high quality network universal and transparent access. ISDN offers two modalities to connect users:
BRI (Basic Rate Interface) or BRA (Basic Rate Access) which provides two B channels (Bearer) and one D channel (Data). The B channels can carry any form of digitized information (data, voice, image and video), while the D channel is primarily intended to carry packet signalling information.
˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙˙ PRI (Primary Rate Interface) or PRA (Primary Rate Access) which provides up to 30 B channels and one D channel.
The transfer rate per B channel is 64 Kbits/s. The transfer rate of the D channel is 16 Kbits/s for the BRA and 2 Mbits/s for the PRA. ISDN supports both circuit switching transmission (public telephone network) and packet switching transmission (X25).
A set of services is supported by ISDN: CLIP: Calling Line Identification Presentation, CLIR: Calling Line Identification Restriction, DDI: Direct Dialling In, MSN: Multiple Subscriber Number, TP:Terminal Portability, AOC: Advise Of Charge, CFB: Call Forwarding Busy, CFNR: Call Forwarding No Reply CFU: Call Forwarding Unconditional, CH: Call Hold COLP: COnnected Line identification, COLR: COnnected Line identification Restriction, CUG: Closer User Group, CW: Call Waiting, SUB: SUBaddressing, UUS: User User Signalling,
Each PNO (Public Network Organization, in most instances still the National PTT) offers a National ISDN with a specific version of the D protocol and therefore a specific set of services of which the complete list is quoted above. Unfortunately the current versions of ISDN are not fully compatible. An International ISDN exists : it is the result of the interconnection of the various national ISDNs but it has been realized with a proprietary protocol D version called code 7 TUP (Telephone User Part) and does not offer the services listed above.
Nevertheless an Euro-ISDN will be a standard for ISDN in Europe and is about to be available. It will offer the first five services.
ATM (Asynchronous Transfer Mode)
ATM is chosen by the ITU-T (formerly CCITT) as standard for the next generation of WAN. In fact this technology can be applied to WAN as well as to LAN. For instance, IP (Internet protocol) will "work" on top of ATM and FDDI and will be connected to ATM. ATM in future will be used as backbone network. ATM is a compromise between the simplicity of circuit switching mode (e.g. ISDN) and the flexibility of the packet switching mode (e.g. X25) and is a so-called Cell Relay technique. The cells are much smaller than with X25 and of fixed size, hence the speed of transmission.
Adaptation layer (AAL): Segmentation and re-assembly of cells. ATM layer Generation and extraction of the cell header and control of the multiplex/demultiplex of cells .
Physical layer: The media can be either shielded or unshielded Twisted Pair or Optical fibre. The transfer rate starts at 25 MBits/s to 155 MBits/s for Twisted Pair and up to several GBits/s for optical fibre.
Internet TCP/IP (Transmission Control Protocol/Internet Protocol)
TCP/IP can be represented in layers like in the OSI model but it does not respect this standard as it was defined before the OSI model. TCP/IP is almost always referred to as a unique form because TCP and IP represent two different layers of protocol. In comparison with the OSI model, IP corresponds to layer 3 (Network) and TCP corresponds to layer 4 (Transport). Thus it can be used either in LAN or WAN environment according to the lower layer used in conjunction with it.
Application: Assurance of data processing (file transfer, mail) TCP (Transmission Control Protocol): Assurance of reliable data transport in a connected mode UDP (User Datagram Protocol): Data transport without control in disconnected mode. IP (Internet Protocol): Fragmentation and re-assembly of datagrams (basic item in IP) and routing of these datagrams.
In LAN environment, TCP/IP is often used with Ethernet.
In WAN, TCP/IP is often used with X25 or Frame relay in lower layers. In higher layers, protocols such as FTAM (File Transfer, Access and Management) for file transfer, X400 for messaging and X500 for Directory services are also used in conjunction with TCP/IP.
TCP/IP often includes at least the three standardized upper layer protocols: Telnet, File Transfer Protocol (FTP) and Simple Mail Transfer Protocol (SMTP).
The RAMA structure for instance is built into layers in the top of TCP/IP. RAMA accepts any LAN with winsockets facilities and any WAN protocols which encapsulate TCP/IP.
X25
The X25 standard met a real success in data transfer. In the beginning public networks were developed such as Telenet in the USA, NPSS in Great Britain, Datapak in Scandinavia and Transpac in France. But beyond a certain threshold it became interesting for important companies to buy and manage their own networks in order to save money.
ISDN
Most Public Network Operators offer National ISDN and are connected to International ISDN (see paragraph 4.4.2). Examples are AT&T and MCI in the USA, BT and MERCURY in Great Britain, ITJ and KDD in Japan, and France Telecom in France.
Until now ISDN did not meet the expected success. One of the reasons is certainly the lack of full compatibility between the different National ISDN. Organizations such as ETSI (European Telecommunication Standards Institute) are working on the harmonization of standards. Unfortunately, in the case of Euro-ISDN, the implementation and wide acceptance of the standards will last several years from 1994 to 1996 or later.
ATM
ATM has reached the maturity necessary to shift from "laboratory experiments" to "field deployment". Manufacturers are releasing commercial ATM products, users are installing first ATM LANs and Network Operators are realizing public ATM networks. In particular, 15 Public Network Operators have developed an ATM pilot in Europe, which after an appropriate consolidation phase probably will evolve to commercial operation.
Internet (TCP/IP)
Internet is an interconnection of hundreds of networks running the TCP/IP protocols to the effect that users of any of the networks can use the network services provided by TCP/IP to reach users on any other networks. In 1994, it encompasses more than 25 million users. Internet started as ARPANET (Defense Advanced Research Project Agency NETwork) in the USA, but now includes such networks as NSFNET (National Science Foundation NETwork) in USA, JANET (Joint Academic Network), RENATER (National Network for Technology, Teaching and Research) in France, EBONE (European IP backBONE) and hundreds of others.
Pricing Volume vs. Connection Time
The communication over Internet is free of charge. Contrary to that communication over other networks are invoiced. The following example tries to give an indication of the communication price according the type of the network. The figures originate from France Telecom services through the products Transpac as X25 Network and Num)ris as ISDN. They relate to file transfer.
France Telecom invoices its ISDN service according to the distance and the duration of the connection whereas it invoices its X25 service mainly according on the basis of data volume and in a lesser way according to duration of the connection.
ISDN X25 (9.6 KBits/s) Local communication (50 Km to 100Km): 100 Kb Compressed still image will take 2 to 15 seconds for 1ECU. A 2Mb Uncompressed still image will take 40 seconds to 4 minutes and 15 seconds and will cost 2ECU.
Communication between France and Great Britain:
100 Kb Compressed still image will take 2 to 15 seconds and cost 5 ECU. A 2Mb Uncompressed still image will take 40 seconds to 4 minutes and 15 seconds and will cost 96 ECU
These figures take into account neither the subscription cost nor the duration of connection separate from data transfer duration. The ISDN service is based more on connect time than the X25 service. During database consultation each minute costs about 0.3 ECU for local communication and 0.6 ECU for European communication using an ISDN service. This is in contrast to the insignificant cost using a X25 service.
Important research work is executed in the scope of the European Union Research & Development Programme in the IVth Framework 1994 -1998, such as ACTS for the Telecommunications experiments.
12. Intellectual Property Rights 14. Conclusion