-How does IP interconnect networks? The IP protocol is actually a set of protocol software composed of software programs, which puts the various "frames" in different formats. Different "frame" unified into "IP datagram" format, this conversion is one of the most important features of the Internet, so that all kinds of computers can be realized on the Internet interoperability, that is, with "openness
-This conversion is one of the most important features of the Internet, enabling all kinds of computers to interoperate on the Internet, which is characterized by "openness.
-What is a "datagram"? And what are its characteristics? The datagram is also a form of packet switching, in which the transmitted data is segmented into "packets" and then transmitted. However, unlike the traditional "connection type" packet switching, it is a "connectionless type", in which each "packet" (subgroup) is treated as a Each "packet" (group) is transmitted as a "separate message", so it is called a "datagram". In this way, there is no need to connect a circuit before starting communication, and each datagram is not necessarily transmitted through the same path, so it is called "connectionless". This is a very important feature that greatly improves the robustness and security of the network.
--Each datagram has a header and the text of the two parts, the header has the destination address and other necessary content, so that each datagram does not go through the same path can accurately reach the destination. At the destination reassembled to restore the original data sent. This requires IP to have the function of packet packing and collection assembly.
--In the actual transmission process, the datagram should also be able to change the length of the datagram according to the size of the packet through the network to change the length of the datagram, IP datagrams can be up to a maximum length of 65535 bytes.
--The IP protocol also has a very important content, that is, to each computer and other devices on the Internet are specified a unique address, called "IP address". This unique address ensures that users can efficiently and easily select the objects they need from millions of computers when operating on a networked computer.
-Now the telecommunication network is converging with the IP network, the new technology based on IP is the hot technology, such as the technology of transmitting voice over IP network (i.e., VoIP) is very popular, and others such as IP over ATM, IP over SDH, IP over WDM, etc., are the focus of the research on IP technology. technology research focus. (IP Global Network)
In today's world to the knowledge-based economy in the process of progress, the application of computer Internet technology has become an important contributing factor to its continuous development to form a new source of impetus to promote the world's rapid economic development. With the in-depth development of national economic informatization process, the whole society on the modernization of communication needs to further increase, a new generation of broadband communication network will become a new generation of telecommunication obvious characteristics, broadband IP network technology came into being.
I. Current Characteristics of IP Services
Initially, IP networks were designed to provide services for a small number of nodes, and to provide resource *** enjoyment and file transfer capability for a small number of users. Currently, IP networks have grown unprecedentedly worldwide, and the way the networks are applied and characterized has changed. Therefore, examining the characteristics of current IP services is the basis for guiding the further development of IP networks.
1. Sharp growth in the number of users
The size of the Internet is now growing about 10% per month, the volume of business doubled every 6-9 months. It is predicted that by the end of 2000, the number of Internet users will reach 300 million to 1 billion. Since 1997, foreign operating companies in the network business has been the data business over the voice business situation; a foreign statistical company analysis, it is expected that in 1998-2008 period of the Internet data traffic in various countries will first exceed the voice, of which North America by the year 2000, the data business will be five times the voice business. 1997 the end of the year, China's data through the user 600,000, at the end of 1998, 220,000 users. At the end of 1997, China's Datatone had 600,000 users, and at the end of 1998, there were 2.2 million, an increase of nearly four times; in June 1999, the number of users had reached 4 million, and it is estimated that by the year 2000, it will be 25 million, and in 2005, it will reach 50 million.
2. Exponential growth in service bandwidth
In addition to the exponential growth in the number of subscribers, the service bandwidth also shows exponential growth. For example, around 1990, the main business is E-mail, the bandwidth is only about 1kbit / s; to 1995, the main business into Web browsing. The number of Web sites in the United States doubled every 57 days, and the current number of WWW sites in China is about 99.06 million. Around 2000, the moving image will become one of the important business, the occupied bandwidth can reach 5Mbit / s. 10 years, the business bandwidth growth can be up to 4 orders of magnitude. All these changes make the bandwidth required for IP services is explosive growth, forming a new period of network bandwidth growth of the main driving force.
3. Integrated Service Content
The TCP/IP protocol was originally designed to provide non-real-time data services. In order to enable IP networks to transmit not only non-real-time
data information, but also real-time multimedia data information, international standardization organizations (e.g., ITU, IETF, etc.) have begun
began to draft and completed a number of standards for real-time IP communications, as well as quality-of-service standards, such as Real-time Transport Protocol (RTP)/RTCP, Real-time Transmission Control Protocol (RTP/RTCP), Real-time Transmission Protocol (RTP/RTCP), and real-time transmission control protocol (RTP/RTCP). RTP/RTCP), Resource Reservation Protocol (RSVP), IP multicast technology, and H.323 recommendations.
Additionally, a lot of research has been done on more advanced quality-of-service assurance technologies such as Diff一Serv(Diff一Serv), etc., and promising results have been achieved.
With the support of these technologies, the Internet has become an important platform for the development of the Internet. With the support of these technologies, the applications and services provided by the Internet will be able to cover the types of services
in the integrated service network.
4. The flow of services is self-similar and send/receive asymmetry
With the increase in the volume of IP network services, the traffic in the network is characterized by self-similarity, i.e., regardless of the number of concurrent business
service streams on a given link, the traffic has the same characteristics. Therefore, to reduce network congestion, IP networks must
have higher average peak-to-load ratios than traditional telecom networks. Similarly, due to the characteristics of the applications on an IP network,
traffic on the network exhibits a significant send/receive asymmetry.
The future development of IP networks and technology
The rapid and sustained growth in the volume of IP services has made the IP protocol gradually become a dominant communications protocol, IP networks in the
future data communications and even the telecommunications industry will occupy an important position. More importantly, the network has been transformed into a commodity, on which a variety of value-added services have huge profit potential. It is reasonable to believe that IP technology is one of the main forces in the future of network integration
it can integrate voice services, data services, image and video services; IP networks may eventually become the technology of choice for the new
generation of telecommunications network infrastructure.
The current IP network and IP technology still has such and such defects, to become the basis of the new generation of telecommunications networks still need to
solve a large number of topics, in all the topics to be solved, network performance is one of the basic conditions, so high-speed broadband IP network
is the prerequisite for the solution to the problem of the development of IP network. Of course, high-speed broadband is relative, the broadband in this paper refers to the rate of 155 Mbit/s
above.
In order to build a high-speed broadband IP network, ITU-T, IETF and ATM Forum and other organizations are working together with a large number of equipment manufacturers
Manufacturers and network service providers *** with the search for solutions to transform the Internet backbone. Generally speaking, the various IP technology solutions
are divided into two ideas: to take the route of IP and ATM combination; or to take the route of optical IP. The former leverages the power of ATM networks to deliver IP based on ATM; the latter is based on the concept of traditional IP networks and leverages the power of optical transmission systems to deliver IP. These technologies are the products of a specific period of time and a specific technological background, and each has its own characteristics and applicable occasions.
Three, IP/ATM broadband network
International ATM to provide IP services have done a lot of research, proposed a series of protocols and standards. These protocols and standards can be summarized into two categories: overlapping types and integrated types.
1. Overlap type
It is recommended to continue to use the existing network to provide IP services, IP network is built on top of the ATM network. the ATM network and the IP network have their own addressing and routing protocols, the use of IP services, ATM user terminals should be both the ATM network address and the IP network address. the routing function of the IP through the ATM routing function to establish a connection, and therefore, in the fat network of the various parts of the network. The IP routing function is established by the ATM routing function, so each node in the fertilizer network should have the mapping function between IP address and ATM address. Some of the IP over ATM technologies that use overlapping types are: the ATM Forum's LAN Emulation and Multi-Protocol over ATM (MPOA); and the IETF's Conventional IP Protocol over ATM Networks (CIPOA).
2. Integration type
Integration type under the network no longer has two layers, the network layer of the ATM switch for IP services is used for IP-specific protocols. Users using IP services only need an IP address, and the switch no longer has the function of translating from an ATM address to an IP address.
Multiprotocol Label Switching (MPLS) is now recognized as a good solution for combining IP and ATM, and ATM's meta-mechanism can very effectively support label switching in MPLS, making it easy for ATM switching to support the forwarding functions of Label Switching Routers (LSRs) in MPLS. MPLS technology is independent of the link layer and can be implemented on both ATM and pure routers.
The characteristics of IP/ATM include: the quality of service characteristics of ATM can be utilized to ensure the quality of service of the network; applicable to a variety of services, the network has a very good expansion performance, the user can be placed on any link in the required capacity; there is a good network traffic management and congestion control performance; applicable to the general IP backbone network.
The shortcomings of IP/ATM include: IP packets need to be mapped into ATM elements, which results in larger transmission overhead and lower transmission efficiency; the need to resolve the contradiction between multiple mappings of IP addresses and ATM addresses and the contradiction between the non-connectivity of IP networks and the connection-oriented characteristics of ATM, which makes the management of the network more complicated; the bandwidth of the IP network realized based on ATM is limited by the ATM network technology itself, which leads to its less suitable for ultra-large IP backbone network (generally considered to be used for ultra-large IP backbone network edge multi-service access).
Four, optical broadband IP network
1. IP over SDH / SONET mode
Can be considered as IP over SDH / SONET is the prototype of optical broadband IP network. IP packets through the use of point-to-point protocol
Programming (PPP) mapped to the SDH / SONET frames, the line rate of the corresponding subgroups by the continuous transmission. line rate for continuous transmission.
The PPP protocol is a simple OSI Layer 2 protocol with a two-byte header, no address information, and just point
to-point order.
The PPP protocol slices IP packets into PPP frames to meet the requirements for mapping to SDR/SONET frame structures.
The implementation of IP over SDH/SONET technology requires high-speed routers and the PPP protocol, using what is still the packet-by-packet forwarding approach of traditional routers
. The basic idea of this method is to separate the routing calculation from the packet forwarding, and to control the packet-by-packet forwarding speed of the router to be comparable to that of Layer 2 switching by using buffering technology,
hardware (chip) fast processing technology, and switching routing technology using the ATM cell switching matrix as the internal architecture of the router
technology. It eliminates the need to utilize
ATM switches on the WAN to establish virtual circuits. Currently, many network equipment companies have launched IP overSDH/SONET technology-based switching
exchange router products.
The features of IP over SDH/SONET are as follows: IP packets are directly mapped to the SDH/SONET frame structure through the PPP protocol, eliminating the intermediate ATM layer, simplifying the IP network architecture, and improving the efficiency of data transmission; IP network technology is built on the SDH/SONET transmission platform, which can easily cross regional and national boundaries. Compatible with a variety of different technologies and standards, to achieve network interconnection; can make full use of the various advantages of SDH / SONET technology, such as automatic protection switching (APS),
to ensure the reliability of the network; conducive to the implementation of IP multicast technology; suitable for large-scale IP backbone network.
The shortcomings of IP over SDH/SONET technology are: SDH is mainly considered as a circuit-switched network indicators, such as
step, self-healing, jitter performance, etc., in the IP network, the requirements of these indicators are not necessarily the same; is not very suitable for the set of data, voice
audio, image and other integrated multi-service platform; IP over SDH/SONET technology is generally available for the network reliability, and can be used for IP multicast.
The IP over SDH/SONET technology is generally capable of Class of Service (CoS),
currently not able to provide better quality of service as IP over ATM technology; lack of circuit emulation service capability; network
expansion is not as flexible as IP over ATM technology.
2. IP over DWDM
From the development trend of optical communication technology, SDH/SONET is inevitably based on Dense Wavelength Division Multiplexing (DWDM) technology, therefore, IP over SDH/SONET will ultimately evolve into IP over DWDM, that is, IP packets are transmitted directly on the optical wave channel.
The use of IP over DWDM technology can reduce redundancy between the layers of the network; reduce the overlap of functions between the layers of SDH/SONET, ATM, IP, etc.; and reduce the cost of equipment operation, maintenance and management. At the same time, due to the elimination of the intermediate ATM layer and SDH/SONET layer, its transmission efficiency is high, which can greatly save the network operating costs, thus indirectly reducing the user to obtain multimedia communication
Communication service costs. This is one of the most direct, simple, and economical IP network architectures for very large IP backbone
The combination of IP and DWDM will result in an all-optical IP network. The all-optical IP network will be optimized according to the characteristics of IP technology and services, thus opening up a new world for IP networks and even telecommunications networks.
IP overDWDM is supposed to be a better solution for broadband IP networks. An all-optical network uses wavelength-selectable optical components at network nodes to separate optical signals of different wavelengths for optical multiplexing and demultiplexing, as well as optical routing and optical
exchanging. DWDM technology is the foundation of an all-optical network. IP over DWDM networking combines the technologies of wavelength routing and IP routing. Wavelength routing provides large-grained multiplexing, while IP routing provides fine-grained multiplexing, and the combination of the two provides an optimized environment for IP applications.
(1) Problems in IP over DWD
IP over DWDM is only beginning to develop, and the ITU and the Optical Interconnect Forum (OIF) are in the process of standardizing it.IP over
The current problems with DWDM include the adaptation of the data network layer to the optical network layer, specification of physical interfaces, and interlayer management.
The choice of frame structure for IP over DWDM is the key to these problems. the DWDM system itself is characterized by service transparency,
which can carry customer layer signals in a variety of formats. The choice of frame structure should take into account several factors: the frame format's contribution to the packing speed and encapsulation efficiency of IP packets, the frame format's contribution to the management functions of the DWDM system, and the types of optical wavelength type converter (OTU) interfaces that can be provided by current DWDM systems.
The currently available choices of IP over DWDM frame structure schemes are basically in two categories: SDH frame format and Gigabit ratio Ethernet frame format.
①Advantages and disadvantages of SDH frame format
The advantages of using SDH frame format include: most of the OTUs of DWDM systems provide SDH interfaces; the frame header of SDH format can contain a large amount of signaling and management information. The signaling can complete the protection switching and other work, and the management information can assist the DWDM system to complete the network management functions.
The limitations of SDH frame format are: since the size of IP packet and the size of SDH frame do not necessarily match, the segmentation and assembly (SAR) processing of SDH frames on the routing switch interface will affect the throughput and performance of the equipment; the use of SDH frame format requires the OTU to provide SDH interface with SDH forwarding and regeneration functions, which increases the cost.
2 Gigabit Ethernet frame format advantages and disadvantages
The advantages of using Gigabit Ethernet frame format are: for DWDM system, should be an open system, through the wavelength converter IP data stream to the fire optical transmission channel, the OTU only need to provide a transparent interface; the current cost is lower; in the routing switch interface does not require SAR operation.
The limitations of using Gigabit Ethernet frame format are: Because Ethernet frame is an asynchronous protocol, it is sensitive to jitter and timing;
Currently, Gigabit Ethernet adopts 10B/8B encoding, which is a little bit less efficient in encapsulation; the Ethernet frame format does not contain management information, which makes it difficult to monitor the performance of the DWDM system; the transmission distance is not as good as that of the SDH frame format method.
The above two types of frame formats have their own advantages and disadvantages, comparatively speaking, the current choice of SDH frame format than Gigabit than the possibility of Ethernet tilt format. At present, the frame format of the problem in the continuous research and development, for example, to solve the SDH frame format for the IP network of some useless overhead bytes and make the simplification; in order to solve the efficiency of the Ethernet frame format of the 10 Gbit / s Ethernet interface research, etc. The ITU also hopes to research a new optical interface, comprehensive consideration of constant bit stream and burst transmission, solve the frame structure and adaptation protocols, proposed optical parameters and other physical interface characteristics, the optical parameters of the interface. The ITU also hopes to develop a new optical interface that comprehensively considers constant bit stream and burst transmission, solves frame structure issues and adaptation protocols, and proposes physical interface characteristics and management issues such as optical parameters.
(2) IP over DWDM Networking
At present, due to the immaturity of OADM and OXC equipment, the development of all-optical network is still at an early stage,
only point-to-point DWDM transmission system. According to the current state of technology, there are two possible ways of IP over DWDM networking.
①Overlay method
Manually configure the wavepath to form the backbone trunk routing of IP, and then carry out the organization of IP routing on it. In DWDM nodes, in addition to the need for wavepath organization and multiplexer/demultiplexer back-to-back connection method, the main use of high-capacity and high-speed IP routing switches for the up and down services of telecommunications signals. From this point of view, the IP over DWDM network should try to form a ring or bus, in order to reduce the inflexibility of the multiplexer/demultiplexer back-to-back connection and the need for optical insertion, deinterpolation, and multiplexing (OADM) equipment, thereby reducing the cost.
②Integration Approach
Using routing switch equipment as the node equipment, the DWDM system is only used as a means of transmission from point to point.
In this way, the routing switch device acts as an intermediate device in the network, which can complete both relay transmission and up and down circuits. At present, the new generation of routing switch products are using routing calculation and packet forwarding separated from the structure, routing calculation capability is greatly enhanced; the use of network topology-oriented forwarding table maintenance, support for large-capacity routing table; packet forwarding mechanism using the switching matrix to achieve non-blocking switching. The port throughput capacity has been roughly close to the line speed in various packet sizes, and the single switching delay has been in tens of microseconds, which is capable of supporting the networking application in this way.
V. Traffic Engineering in Broadband IP Networks
In the early 1990s, when IP networks consisted mainly of links below 155 Mbit/S, traffic engineering was realized primarily through the use of routing metrics. However, as networks grew in size and complexity, metric-based traffic control became
more and more complex, to the point that when adjustments were made to a portion of the network's metrics, it became increasingly difficult to determine the impact of the adjustment on the rest of the network, making it difficult to achieve a comprehensive and efficient use of bandwidth across the entire network.
In an IP network based on ATM PVC links, it is possible to use ATM traffic engineering to partially meet service requirements. However, PVC links used as backups must be configured and installed into ATM switches in advance, and it is difficult to design backup PVCs that are similar to the inherent recoverability of IP due to the uncertainty of the failure node.
The advent of MPLS has provided a promising solution to the problem of traffic engineering in IP networks. In broadband IP networks,
especially in all-optical IP networks, MPLS is even now the only effective solution to the traffic engineering problem. Traffic engineering is essentially
mapping services to a physical topology, and MPLS does this by creating labeled switching paths between inputs and outputs.MPLS can compute full or partial labeled switching paths offline, and it can use dynamic protocols to install
these paths. In the future, MPLS can support constraint-based routing, where the network itself participates in the calculation of labeled exchange paths, reducing the stress and lack of
manual participation, and accelerating traffic engineering
reaction to and recovery from failures through dynamic routing information dissemination via Intradomain Routing Protocols (IGPs).
VI. Quality of Service in Broadband IP Networks
Quality of service is a constant topic in the development of IP networks, and a relatively "fragile" aspect of IP networks. Although increasing bandwidth sufficiently
to exceed service demand can effectively address QoS, device port buffer capacity limitations and the constant
appearance of new applications make the bandwidth growth almost never be able to satisfy the demand for all services, so the broadband IP network must still deal with
QoS issues.
There are currently three stages of end-to-end quality of service levels in IP networks: Best-Effort,
Differen-tiated, and Assured. Best-Effort is the current status of service quality in most IP networks, and Differen-tiated and Assured services are under constant development. Differen-tiated service is a soft QoS concept that provides statistically significant prioritization, while guaranteed service must reserve defined network resources for specific services.
The IETF currently investigates two main models for addressing quality of service, namely Integrated Services (Int-Serv) and Classified
Services (Diff-sarv). The former leads to scalability problems due to flow-oriented solutions, and many people are skeptical about the practical operability of utilizing full signaling technology to achieve QoS; the latter does not address end-to-Rui QoS in isolation, but rather classifies services in the context of domains, and is supported by QoS policies of the devices, such as queue management in the network, rate control based on the leaky bucket principle, and congestion management based on packet loss policies. management based on packet loss policy, etc., to realize the quality of service assurance. Among them, the combination of connection-oriented capability through MPLS and simple signaling technology through Diff-serv may provide a better solution to address the quality of service in IP networks.
VII. Self-healing technology in broadband IP network
The huge bandwidth carrying a large number of services makes the reliability of broadband IP network is more important, at present, due to the DWDM system is only commercially available point-to-point system, so the self-healing of the IP over DWDM network protection can only be used in the optical layer of the optical fiber protection of the 1 +1. At the IP layer, the IP network is inherently self-healing when dynamic routing protocols are used. These two types of protection have different effects: the protection time at the optical layer is in the milliseconds; while at the IP layer, since its self-healing function is realized by re-routing, the length of the protection time depends on the time required for the routing protocol to discover the link state change and the time for the routing computation to
re-converge (ten generally in the tens of seconds or so).
For most common IP applications, protection at the IP layer is sufficient because the application bottleneck is typically at the server, not the
network. But for real-time IP-based applications, seconds can affect service quality. Currently based on MPLS traffic engineering fast change routing characteristics of the IP network self-healing protection has been basically realized to the Is about the path switching.
Eight, foreign broadband IP network construction dynamics
1. ultra-high-performance backbone network service plan
1993, the U.S. National Science Foundation (NSF) began to recognize the need for a higher performance than the then Internet,
faster network to support research. At the same time, the federal government's ongoing High-Performance Computing and Communications (HPCC)
program also required a high-performance network. As a result, NSF decided to implement the Very High Performance Backbone Network Service (VBNS)
program.
In April 1995, NSF and MCI jointly released the vBNS program, which is a five-year program under NSF's responsibility to build a nationwide backbone network with a bandwidth of 622 Mbit/s using MCI's fiber-optic cable network and advanced switching technology to provide a broadband network for scientific research and network application studies.
The most important feature of vBNS was the use of the then advanced ATM technology and SONET transmission technology to build a broadband IP network over the fiber optic cable network by IP
over ATM method. vBNS has a backbone network with a connectivity bandwidth of 622Mbit/s, and is planned to be upgraded to 2.5Gbit/s in 1999. vBNS has a backbone sink on the backbone network, and a backbone sink on the backbone network. backbone with backbone sinks (PoPs), where users are individually connected to the
backbone through the nearest PoP with an access rate of 622 Mbit/s.
vBNS is a broadband network designed for scientific and research purposes, and initially focuses on providing high-speed interconnections between the network and supercomputing centers and NSF-designated access points to the network.
vBNS is designed with 12 The VBNS is designed with 12 PoPs and currently connects five supercomputing centers and 17 universities, with plans to allow an additional 47 universities to connect to the vBNS.
The vBNS is not intended for commercial applications, but only to support connected scientific and research institutions and universities in a wide range of areas including high-performance networked computation,
broadband multimedia network applications, advanced routing technologies, multicast technologies, quality of service and its control, and next-generation interconnections. control technologies, and next-generation interconnection
network protocols (IPv6).
2. Second-generation Internet program
Another advanced network test project in the world is the second-generation Internet (Internet 2) program jointly proposed by more than 80 universities in the United States. The University Consortium for Advanced Internet Development (UCAID) was established to collaborate on next-generation Internet technologies and broadband network applications. Currently, more than 130 universities have participated in the Internet 2 program.
The Internet 2 program has not been fully implemented since it was proposed. Until 1996, the U.S. government put forward the Next Generation Internet (NGI) initiative, strongly support the development of a new generation of broadband network technology, the next generation of interconnection networks as the future of the National Information Infrastructure (NII), the next generation of the Internet performance should be 100 to 1,000 times higher than the current, and can be seamlessly connected to a variety of commercially operated networks. To this end, the United States Government has put aside $300 million to support the construction of a broadband experimental backbone network and the development of new-generation network technologies and broadband applications. With the support of the U.S. government, in 1998, UCAID proposed the Abilene program to build a high-speed national backbone network in support of the Internet 2 initiative through a partnership with Cisco, Nortel Networks and Qwest. Cisco provided high-performance routing and switching equipment, Nortel Networks provided network engineering and services, and Qwest provided the fiber optic cable for the backbone. With this, the Internet 2 program is now truly fully operational.
The backbone of the Abilene program uses advanced IP overSONET technology, eliminating ATM equipment and transmitting IP packets directly over the SONET/SDH network. The bandwidth of the backbone network is 2.5 Gbit/s. It is planned to set up one gigabit backbone convergence point in the country, and 64 members will be connected at the end of 1999. Universities participating in Internet 2 will be connected to the backbone network at three rates of 155, 622 Mbit/s and 2.5 Gbit/s through nearby access points, realizing the interconnection of broadband networks with a gigabit ratio.
Abilene plans to upgrade the backbone to 9.6 Gbit/s in the future, and construction of the backbone began in early 1999.
In Abilene's broadband backbone. With the support of Abilene's broadband backbone network, Internet 2 will carry out research and experiments on various broadband network technologies and applications, of which the main network technology is quality of service control technology. To this end, Internet 2 has established the Qbone network to develop a unified quality of service control technology, forming an international standard, which will have a huge impact on the next generation of the Internet, the new generation of network communication equipment industry and future broadband network applications.
The other main purpose of Internet 2 is to support the research, development, and testing of broadband multimedia network applications, mainly collaborative design, collaborative experimentation, distance education, telemedicine, broadband conference television. Video-on-demand, video multicasting, virtual reality, remote operation of scientific instruments, etc. These applications will become the main applications of the future next-generation Internet.
3. Advanced Network Third Generation Program
Early in 1998, the Canadian government put forward the Canadian Advanced Network Second Generation (CANet 2) program, in conjunction with the U.S. Internet 2 program, which plans to use the technology of IP over SONET/SDH to set up a high-speed backbone test network across the country. To this end, a special non-profit company was set up to be responsible for the planning, construction, operation, management and maintenance of the CANet 2 network. In September 1998, the Canadian government also substantially upgraded the CANet 2 program, and put forward the CANet 3 program, which is the first one in the world to adopt the most advanced all-optical network technology on the basis of CANet 2 to establish the the world's widest national high-speed backbone test network.
CANet 3 uses DWDM technology to simultaneously transmit multiple optical signals over a pair of optical fibers, increasing the transmission bandwidth of optical fibers by ten or even dozens of times. In addition, CANet 3 also transmits IP packets directly over the DWDM fiber network in SDH frame format, which greatly improves transmission efficiency and reduces network construction and operating costs.
The CANet 3 backbone network runs from Vancouver in the west to Halifax in the east, passing through Chicago, USA, and connecting to STAR TAP, the Internet's aggregation point. 13 gigabit backbone sinks are planned for CANet 3, and each access network is connected to the backbone through the nearest sinks at the rate of 155, 622 Mbit/s, and 2.5 Gbit/s. Similar to Internet 2, CANet 3 is also connected to the Internet through the Internet's Internet Protocol (IP), which is the most widely used IP network in the world, and the most widely used IP network in the world. Similar to Internet 2, the main objective of CANet 3 is to support Canadian research institutions and universities in their research on next-generation Internet technologies and future broadband network applications. In addition to research on quality of service control technologies and broadband multimedia network applications, research will also be conducted on network technologies such as the combination of high-performance routing switches and DWDM, network self-healing recovery technologies, and traffic engineering.
In addition, the international network proposed to use IP over DWDM technology are Sprint, MCI, KDD's KTH21, the largest trans-European fiber optic network GTS and so on.
Nine, China's broadband IP network outlook
Emerging China Network Communications Ltd. plans to implement China's high-speed interconnection Internet demonstration project, which will use IP over DWDM technology to build a new generation of high-speed broadband network, taking the first step in the construction of China's broadband IP network, and its main business is aimed at providing broadband wholesale services, broadband access services, IP phone Its main business aims to provide broadband wholesale services, broadband access services, IP telephony services and various IP services. Previously, China Telecom also adopted IP over SDH technology in its IP network to improve the transmission rate and capacity of the network. Other operating companies are planning solutions for broadbandization of their networks. With the development of information technology and the further popularization of IP applications, the volume of information will grow. It is believed that broadband IP network with its high-speed, broadband, flexible and convenient advantages continue to be applied in China, and its development prospects are very broad.