The Next Generation Internet

The mind of the beginner is empty, free of the habits of the expert, ready to accept, to doubt, and open to all the possibilities.-- Shunryu Suzuki --

[HOME]

[Security Architecture]

W3C Activities

[Professional XML]

E-commerce - Case Studies

[Feedback Form]

Applications

Protocols Networks Projects

Mapped the status of Hong Kong's Internet development to the framework I find that Hong Kong is inactive in partnering / joining alliance. We are also lack of project initiatives and advanced applications development, like e-Learning. If Hong Kong's role as an "middle-man" between the West and China. Then in the Internet Age we may focus on middleware development, maintaining a platform / environment for data transactions and information interchange. Along with the physical goods that are moving, there are huge amount of information / records moving too. In this case, Hong Kong's identity should become an electronic-middleman, or "e-middleman".

Applications

E-Commerce, -Government, -Learning

E-Government Governments are using Internet to enhance interactions with citizens. The Internet-based systems also allow better communications among government agencies.

One of the immediate benefits of transforming E-government is cost reduction. Forrester, in the report "Sizing US eGovernment" 2000, forecasted that the United States alone, federal, state, and local governments would collect over US$600 billion22 (including government fees and tax) via the Internet by 2006. On the servicing side, governments are looking to the development of e-business applications to improve service delivery and efficiency. Apart from the back-end process re-engineering, there will also be front-end integration with private sector services. During the course of transformation (E-government), governance, policy, regulations, and infrastructure will change to meet the requirements. This provide an environment for the E-Commerce to grow.

HK SAR Government Information Technology Strategy - Digital 21

E-Commerce Internet break the limitation of time and distance in doing business. Innovative business models / service delivery models will emerge to boost the economic growth.

E-Learning Internet-based digital learning and research environments not only can enhance the classroom learning but also provide capabilities for life-long learning by anyone wishing to continue his or her education. In the Age of Knowledge we are living this is a crucial factor for a country to be able to grow persistently.

In Hong Kong, Hong Kong Education City, a project launched by the Education Department of the Hong Kong SAR Government, is to lead, to serve primary and secondary schools, teachers, students, parents, and the public. Its goal is to promote Quality Education and IT Culture for life-long and life-wide learning.

What is the performance of Hong Kong in the "E-xxx" area? What do you think?

Digital-Library, -Museum

Digital-Library Todays computer systems are powerful enough to handle not only text, images, video, and audio but also simulations, animations, and digital objects. Digital libraries are being designed to hold all these contents. In the next generation of Internet, protocols are more efficient and the network is fast. User connected to the digital library can interact with the contents to learn what they would like / need to learn.

Digital-Museum Like the digital library, digital museum is not simply collections of art and historic artifacts. Advanced Internet-based virtual-reality technology provides user an experience of, like traveling back to the past, historic sites visit. Because this information is digital, it can be cross-linked. Therefore, applications related to digital objects should provide not only the objects' data but also related metadata that can be searched and cross-linked.

Digital ibraries and digital museums are important parts of constituting the E-Learning environment.

Tele-Medicine, -Vision, -Collaboration

Tele-Medicine Medical researchers are developing numerous advanced, network-based applications utilizing medical software. For example, medical imaging is a fast-growing application area that enables radically new types of health care. The next generation of Internet will link hospitals, clinics, medical schools, universities, and research centres in a secure manner for data and information transmission, remote diagnostics, share of instrumentation, development of new pharmaceutical products, enhanced administrative and patient care processes, new procedures for health maintenance, and professional skills training.

Tele-Vision Nowadays, powerful data visualization software allows us to interact dynamically with the models from the computer simulations. For example, data visualization is used extensively in the network-based computational molecular biology and computer aided design of car produce. The promise of these techniques relies on the ability to access extremely large amounts of stored data and to link that information to displays over a network. The more information that can be streamed, the more detail that can be imaged.

Tele-Collaboration Internet breaks the boundary of the countries. With the collaborative software, workers spreaded in different geographical regions are able to work in synchronical, real-time mode.

On-demand Movie, Music, Entertainment

On-demand Movie, Music, Entertainment The next generation of Internet will provide interactive digital entertainment that linked to many types of digital data. The content will have dynamic links to many types of digital information, includes text, images, separate video, and audio tracks. For the business purposes, it will have links to different types of transaction-processing systems. Because digital video streams are latency intolerant, they require more than bandwidth. They require that network resources be reserved to ensure that the streams are continuous.

Transforming the traditional government to E-Government is a strategy of developing E-Commerce. In the Age of Information, life-long education is important for the development of a knowledge society. So we need to build the platforms for E-Learing and look into the facility of digital-library.

These applications run in the N-tier client/server environments. The followings are required to make things happen:

Protocols

IP Addressing

IPv4 is currently used in the Internet. The length of the header is 20 bytes. IPv4 provides various Network Level services:

  • Addresses - 32-bit source IP address and destination IP address using Dotted Decimal Notation (DDN).
  • Protocol type
  • Data unit ID - unique integer that identifies the datagram, allows the destination to collect datagram fragments into an integral datagram.
  • Type of service - i.e., expected delay characteristics, expected reliability of path, etc
  • Time-to-live - determine the PDU's lifetime in the interconnected system.
  • Options - parameters to specify security, timestamps, and special routing.
  • Header checksum - a two-octet field used by IP to determine PDU integrity (no corrective action is supported).

IPv6 has been proposed as a replacement for IPv4. IPv6 uses 128-bit addresses. There is a 24-bit flow label field among flows to indicate different classes of services for the quality of service applications. Network security has been enhanced, IPv6 has developed the concepts of IP-level encryption and address authentication.

IPv6 uses colons to separate the address into eight 16-bit segments. It supports three type of addresses in the target field:

  1. Unicast (source field must be a unicast address)
  2. Multicast
  3. Anycast

The header for IPv6 is 40 bytes fixed. Because the source and destination addresses are longer, the rest of the header has to be simpler so that the protocol is simpler as well. Many of the IPv4 header functions are carries as extension headers. The fields are as follows:

  • Version - the version of the protocol, now 6 (4-bits).
  • Priority - 4-bit nibble replaces the functions of the Precedence field in IPv4. The lower the priority value, the more willing the source is to have a router discard the PDU.
  • Flow label - 3-byte field is used to request special handling of the PDU in the router(s). The router that recognize the flow label can avoid routing tasks at the network layer (e.g., looking up a router table) by simply following the previous calculations/assignments made in forwarding previous PDUs in this flow. This is the concept of tag switching / netflows / MPLS. All PDUs with the same nonzero flow label must have the same destination address, Hop-by-Hop Options header, Routing header, and source address content.
  • Payload length - 2-byte unsigned integer field specifies the length (in octets) of the PDU after the IPv6 header.
  • Next Header values - 1-byte field performs the same function as the IPv4 Protocol field.
  • Hop Limit - the maximum number of nodes that may forward the PDU. An octet, can up to the integer 255.
  • Source/Target Address - 16-octet address of the sender/recipient of the PDU. (If a Routing header is present, it may not be the ultimate destination.)

Middleware

The function of the Middleware is to match the requirements of applications to the resources provided by the network. They are:

  • interlinking/brokering tasks
  • processes
  • services
that designed to work between the applications and the infrastructure.

At the application level, middleware runs on both the client and servers sides. They are found in the transport stacks, network operating systems, and service-specific layer. Middleware opens communication pipes for the client and server to exchange data. Examples of pipes are RPCs, ORBs, MOMs, SSL and LDAP. There are also service-specific pipes, for instance,

  • Database-specific middleware: ODBC, JDBC, SQLJ, EDA/SQL, OLE DB, OQL, and Oracle SQL*Net.
  • OLTP-specific middleware: Tuxedo's ATMI and /WS, Encina's Transactional RPC, X/Open's TxRPC and XATMI, CORBA's OTS, and Microsoft's DTC and TIP.
  • Groupware-specific middleware: MAPI, VIM, JavaMail, SMTP, Web/NNTP, S/MIME, POP3, IMAP, Workflow, and Lotus Notes calls.
  • Object-specific middleware: OMG's CORBA/IIOP, Microsoft's COM+, and JavaSoft's RMI-over-IIOP.
  • Internet-specific middleware: HTTP, CGI, XML, and SET.
  • System management-specific middleware: SNMP, CMIP, RMON, DMTF, WFM, JMAPI, WEBEM, and ORBs.
In the N-tier environment, as a platform, middleware runs server-side components on the application server provides loading balancing, transaction logging, system monitoring, and security control services.

The applications for the next generation Internet need fast network and intelligent resources management. We should develop middleware platforms base on new Internet technologies to facilitate E-commerce as Hong Kong already built a very good IS infrastructure and be the hub of transportation in Asia.

Routing

Although routers work well for traditional data applications, new broadband video and multimedia applications need different forwarding treatment, higher throughput, and tighter QoS control. The connectivity of routers is from the Network Layer. The most common network layer protocol is IP. Two types of routing protocols are used to create routing tables:

  1. Interior Gateway Protocols (IGP) - Examples, Router Information Protocol (RIP) and the Open Shortest Path First (OSPF) protocol used in the dynamic updates of routing-table information. OSPF is an link-state internal gateway protocol developed by the IETF. Compared to distance-vector protocol, OSPF supports user-definable, least-cost, multipath routing. Each OSPF router calculates its own route using the topology database. This information is periodically proliferated to all routers in the same autonomous system. OSPF does the following:
    • Authenticates routing update information to ascertain it is valid
    • Converges rapidly on network topology changes
    • Is resilient to routing loops
    • Supports load balancing across multiple communication links/services because OSPF can store multiple routes for a destination
    • Supports Type of Service routing, such as link bandwidth or expected link latency, although this is almost never implemented
  2. Exterior Gateway Protocols (EGP) - Example, Border Gateway Protocol (BGP). BGP is a distance-vector exterior routing protocol developed by IETF replacement for EGP. BGP replaced EGP by providing additional features and eliminating some of EGP's limitations. The enhanced capabilities of BGP include (1) support for policy-based routing and (2) use of an authentication technique to guard against unauthorized updates to routing tables.
All IP datagram contain a source and a destination address. Routing in the Internet uses the destination address as the primary means for selecting routes. In the next generation Internet, the routing decision will take into consideration the type of service, network congestion, or network costs.

Currently, IPv6 address allocation architecture (RFC 1887) is similar, if not identical, to the IPv4 address allocation architecture. Both are based on hierarchical routing and specifically, on classless interdomain routing (CIDR) (RFC 1518, 1519). In CIDR, a network in the Internet can be represented by a prefix and a number representing the subnet mask length. CIDR's allocation scheme creates the ability for address aggregation, that is, the practice of aggregating a contiguous block of addresses into a single routing table entry. As a result, the IPv6 routing architecture does not offer any significant improvements over the IPv4's.

Networks

ATM

ATM over SONET/SDH Advanced IP networks are deployed as IP over ATM over SONET/SDH (Synchronous Optical Network / Synchronous Digital Hierarchy). SONET/SDH, as well as optical layers, is traditionally regarded as core transport layers. The SONET standard defines electrically equivalent sychronous transport signals (STS) for fibre-based transmission. The basic data transmission unit is STS-1, which has a bit-rate of 51.84 Mbps. Higher rates of speed are multiples of STS-1. SONET was developed in part to allow for using fibre for long-distance transmission while maintaining an ability to interconnect with existing copper infrastructure. Add-drop multiplexers allow for provisioning circuits within a SONET environment. The Optical Carrier (OC) levels are based on multiples of Optical Carrier 1 (OC1), which is in the speed of 51.84Mbps, for example, OC3 is 155.52 Mbps, OC12 is 622.08 Mbps, and OC192 is 9.95 Gbps. An OC3c designation indicates that the link is implemented as a single concatenated channel, not merely multiplexed STS-1.

Real-Time Protocols Three primary protocols have been developed for real-time QoS data over IP. The fourth is related to stream multimedia content:

  1. Real-Time Transport (RTP) - "end-to-end" protocol utilizing existing Transport layers for real-time applications.
  2. Real-Time Control Protocol (RTCP) - provides feedback on total performance and the quality of the data transmitted so that modifications can be made.
  3. Real-Time Streamming Protocol (RTSP) - A Transport Layer protocol designed specifically for controlling the transmission of audio/video over the Internet.
  4. Resource Reservation Protocol (RSVP) - A multicast-capable resource setup (signaling) protocol primarily designed for IP. RSVP is a general-purpose signaling protocol and can be used to map resource reservations to ATM signaling messages.

Dense Wave Division Multiplexing (DWDM)

Wave division multiplexing (WDM) systems are optical transmission systems in which a single fibre carries multiple optical signals in parallel. In the early 1990s, WDM systems standardized on 1550 nm only support 2 to 4 channels. Later, when a higher level was achieved, 16 channels, WDM was renamed DWDM. Time division multiplexing (TDM) systems expand capacity by increasing the speed of a single optical carrier, such as migrating from OC3 to OC12. DWDM provides more capacity than TDM by multiplexing several wavelengths in parallel.

DWDM can simultaneously carry a wide variety of different protocols. This allows the network infrastructure to seamlessly support any combination of SONET, ATM, and IP traffic at the same time and in parallel. As the Internet evolves, new services and protocols are continually being created. They can be transported on new wavelengths, utilizing DWDM bit-rate performance and format independence. For format independence, DWDM provides a single facility to carry multiple types of Data Link and Network layer protocols simultaneously.

DWDM also supports direct transport of data. For example, an ATM signal coming an ATM switch can transported directly as a wavelength on an optical networking infrastructure without traversing intermdiate SONET multiplexers. The "self-healing" capability of SONET is provided by MPLS's traffic management and control functions.

Currently, DWDM is used in the long-haul transport. This takes advantage of inline optical amplifiers to simultaneously amplify all channels over long distances much more ecconmically than traditional TDM. An emerging trend is the migration of DWDM from long-haul applications into regional and metro transport and to corporate WANs. Access is provided through the optimization of architecture based on optical transport in metropolitan applications. In China, some large cities, are undergoing this changes.

Multiprotocol Label Switching (MPLS)

MPLS is developed by the IETF. MPLS creates a set of protocols that enables traffic management. This is achieved by creating an association between packet's destination and a label. A label is a short, fixed-length identifier that is used to forward packets. An MPLS device will use the labels to make forwarding decisions on packets and will usually replace the label in the packet with some new value before hopping. MPLS devices run standard IP routing protocols to determine how to establish their label forwarding table, in the same way they would create routing tables. In this case, MPLS devices blur the lines between traditional routers and switches. IP routers can support MPLS just as MPLS allows unmodified ATM hardware to implement the protocols.

Broadband

The Broadband Internet Connections

Third Generation Wireless Systems

Evolution The cellular concept originated at Bell Laboratories in 1947. All mobile radio systems use cellular technology, includes Personal communication Services (PCS) and Personal Communications Network (PCN) systems. First Generation cellular 1G is a hybrid of analog voice channels and digital control channels. The analog voice channels used FM and the digital control channels used FSK Frequency Shift Keying modulation. Because of the limited transferring rates the systems were lack of advanced authentication and encryption features. The Second Generation cellular and PCS/PCN 2G use digital radio channels for both voice and control. 2G use multiple access technologies to allow more users to share individual radio channels or use narrow channels to allow more radio channels into a limited amount of radio band. There are three basic types:

  1. FDMA Frequency Division Multiple Access - reduce the RF channel bandwidth
  2. TDMA (IS-136) Time Division Multiple Access - share a radio channel by assigning users to brief time slot
  3. CDMA (IS-95) Code Division Multiple Access - divide a wide RF channel into many different coded channels
The all digital systems can provide: short messaging service, web browsing, and enhanced authentication and voice privacy. For example, GSM Global System for Mobile communications can provide voice, low-speed data, and short messaging services (SMS). GSM uses time division multiplexing (TDM) to share one modulated carrier frequency radio waveform among 8 to 16 conversations. 2.5G is the enhancement of 2G but does not quite satisfy the third generation wireless requirements. The 2.5G uses improved digital radio technology to increase the data transmission rates and by using packet-based technologies. There are mainly three technologies used:
  1. GPRS General Packet Radio Service - GPRS dynamically assigns time slots on GSM radio channels to allow quick and efficient transfer small packets of data. GPRS uses the same radio channel structure as the GSM. The maximum data transmission rate is 171.2 kbps. GPRS support point-to-point and point-to-multipoint communications.
  2. HSCSD High-Speed Circuit-Switched Data - by combining more than one traffic channel (TCH/F) of GSM to transfer data. The maximum data rate is 64 kbps compared to the maximum rate of 9.6 kbps of GSM. The speed is possible to increase by a factor of 2 - 4 through the added use of GSM data compression technology.
  3. EDGE Enhanced Data Rates for Global Evolution - is an evolved version of GSM that uses the new 8-level Phase Shift Keying 8PSK (or QPSK) to provide three times the amount of information that is transferred by a standard 2-level GMSK signal used by the first generation of GSM. This results in a radio-channel data-transmission rate of 604.8 kbps and a net maximum delivered data transmission rate of about 474 kbps. EDGE Compact is a version of EDGE that allows the close packing of GSM radio channel frequencies to allow an overlay of GSM into other systems, such as IS-136 TDMA, with a minimum loss of existing channel frequencies.

3G Third Generation is called Universal Mobile Telecommunications Systems UMTS. The requirements for the 3G were first defined in the International Mobile Telecommunications 2000 IMT-2000 system. IMT-2000 main requirements are:

  • High-speed broadband data services - In-building 2 Mbps, Out-door 384 kbps, Wide-area 144 kbps.
  • Multimedia support - delivers different types of information such as voice, data, and video simultaneously or separately.
  • Backward compatible with 2G - allows user to roam globally (different frequency bands) and be able to hand off to 2G systems.
  • Improved system efficiency - capacity increases by allowing more users share the same radio channel spectrum. The third generation UMTS systems increase the overall efficiency by 2 - 4 times compared to 2G systems.
Three are three different system specifications for 3G:
  • WCDMA Wideband Code Division Multiple Access - uses direct Sequence Code division Multiple Access DS-CDMA, efficient QPSK modulation, Paired Frequency Division Duplex FDD RF channels, and variable bandwidth control.
  • TD/CDMA Time Division Duplex/Code Division Multiple Access - uses Time Division Duplex TDD that does not require the paired frequencies of WCDMA. In some countries, paired frequencies are not available. TDD is a process of allowing two way communications by timesharing. The TD/CDMA uses the same DS-CDMA channel-coding to maintain compatibility with the WCDMA system. It is anticipated that TD/CDMA will be used for indoor environment and WCDMA systems will be used for wide area operation.
  • CDMA2000 - is an evolved version of the 2G IS-95 CDMA system. It combines multiple IS-95 radio channels (called multi-carrier transmission) with enhanced packet transmission protocols. These multiples are 3, 6, 9, 12 times the standard 1.25 MHz wide bandwidth.

The 2.5G may meet the price-performance needs of most subscribers sufficiently well that it may effectively compete with 3G systems. For example, EDGE systems are sometimes called GSM384 indicate that it is capable of the UMTS outdoor transmission requirement of 384 kbps.

In Japan, DoCoMo is very success in formulating new business models without using the 3G technologies. Anything we can think of to do on the 2.5G in Hong Kong?

Projects

Internet2

Internet2 is a consortium in United State being led by over 190 universities working in partnership with industry and government to develop and deploy advanced network applications and technologies, accelerating the creation of tomorrow's Internet. Internet2 is recreating the partnership among academia, industry and government that fostered today's Internet in its infancy. Through Internet2 Working Groups and initiatives, Internet2 members are collaborating on:

Asia Pacific Advanced Networking (APAN)

Asia Pacific Advanced Networking APAN Consortium was formed under a Memorandum of Understanding in Jun 1997 to promote advanced research in networking technologies and the development of high-performance broadband applications. There are many advanced networking inititatives in Asia Pacific, and a number of them are interconnected through the APAN. The primary members are:

HARNET is the wide area network which links up the campus networks of the seven tertiary institutions in Hong Kong. HARNET is under the management of the Joint Universities Computer Centre (JUCC), a joint organization of the Computer Centres of the seven member institutions: (1) The University of Hong Kong, (2) The Chinese University of Hong Kong, (3) The Hong Kong Polytechnic University, (4) City University of Hong Kong, (5) Hong Kong Baptist University, (6) The Hong Kong University of Science and Technology, and (7) Lingnan College of Hong Kong.

Resources link

Abbreviations

    Reference:
  1. Joel Mambretti & Andrew Schmidt (1999) Next-Generation Internet: Creating Advanced Networks and Services, Wiley Computer Publish.
  2. Robert Orfali, Dan Harkey, and Jeri Edwards (1999) Client/Server Survival Guide, 3rd edition, Wiley.
  3. Lawrence Harte, Roman Kitka, and Richard Levine (2002) 3G Wireless Demystified, McGraw-Hill.