The chip speed doubles every 18 months, while the Internet traffic doubles every 6 months. As the hub of the Internet,
routers are developing in the three directions of faster speed, better service quality, and easier integrated management.
faster
In the traditional sense,
routers are usually regarded as the bottleneck of network speed. When the speed of the local area network has already reached hundreds of megabytes, the processing speed of the router can only reach a bit rate of tens of megabytes. In recent years, with the explosive growth of Internet users, everyone's research on routers has also focused on improving the processing speed of routers. From 1996 to 1997, a number of innovative small companies appeared in the United States, such as Nexabit, Juniper, Avici, etc., which increased the processing speed of routers to the extreme, and launched Gigabit routers in a short period of time. Even Cisco can only beat it in terms of speed. Because these high-speed routers without exception have introduced a switching structure, they are also called GSR-Gigabit Switch Routers. The optical interface speed of these routers also quickly increased from OC-12 (622Mbps) to OC-48 (2.5Gbps), and then to OC-192 (10Gbps). This speed has already left ATM switches far behind. Since then, the irreplaceable position of ATM in the core network has completely shaken. The protracted IP and ATM technology dispute finally ended with the overwhelming advantage of IP. However, the increase in the speed of IP routers is directly due to the concept and technology of ATM. Many new concepts and technologies proposed in the IP field are also directly or indirectly derived from ATM. The two excellent technologies have gradually begun to merge. . In fact, the technicians of many companies engaged in the research and development of high-speed IP routers are those who studied ATM technology in the past. Specifically, the rapid increase in the speed of IP routers comes from the following four technological advances:
①Hardware architecture
The hardware architecture of the router has roughly undergone 6 changes, from the earliest single bus and single CPU structure to single bus, multiple CPUs and then multiple buses and multiple CPUs. Up to now, high-speed IP routers have mostly learned from the ATM method and adopted a crossbar switch to achieve wire-speed non-blocking interconnection between ports. The technology of the high-speed crossbar switch has been very mature and has been widely used in ATM and high-speed parallel computers. The speed of the high-speed crossbar switch that can be directly purchased on the market is as high as 50Gbps. With the introduction of high-speed crossbar switches, some corresponding technical issues have been introduced, especially for IP multicast, broadcast and quality of service (QoS), using mature scheduling strategies and algorithms, these issues have been well resolved .
②ASIC technology
In recent years, due to cost and performance considerations, ASICs have become more and more widely used, and they are almost certainly called ASICs. To greatly increase the speed in a router, the first thing that comes to mind is also an ASIC. Some use ASICs for packet forwarding, and some use ASICs to check routing, and ASIC chips for IPv4 routing are already on the market. In the trend of ASIC development and mass application, there is a trend worth noting. This is the emergence of the so-called programmable ASIC, which is probably a result of the rapid changes in the network itself. Due to the considerable investment in ASIC design and production, generally speaking, ASICs are only used in fully standardized processes, and the network structure and protocols change quite quickly. Therefore, in the field of network equipment, there has been peculiar Programming ASIC". Currently, there are two types of so-called "programmable ASICs". One type is represented by 3COM's FIRE (Flexible Intelligent RouTIng Engine) chip. This ASIC chip has a CPU embedded in it, so it has a certain degree of flexibility; the other type is represented by Vertex Networks' HISC dedicated chip. The chip is a CPU dedicated to communication protocol processing. Its architecture is designed to specifically adapt to protocol processing. By rewriting the microcode, this dedicated chip can be capable of processing different protocols to adapt to similar changes from IPv4 to IPv6.
③3 layer exchange
This is a revolutionary breakthrough in the protocol processing process, and it is also the source of the names of GSR and TSR. Since the little-known Ipsilon company launched the "route once, then switch" IP Switch technology in 1994, major companies have launched their own proprietary layer 3 switching technology. Such as Cisco's Tag Switch, 3Com's Label Switch, etc. Combining the advantages of these proprietary technologies, the IETF finally launched a multi-protocol label switching (MPLS) with superior performance in 1998.
④IP over SDH, IP over DWDM
The technological progress in this area is entirely derived from the progress of optical fiber communication technology. As the core position of IP is gradually recognized, IP over ATM and then ATM over SDH are replaced by IP over SDH. SDH uses time division multiplexing to carry multiple channels of data. Therefore, a large number of multiplexer cross-connectors are used in the core network. DWDM (Dense Wavelength Division Multiplexing) makes it possible to transmit multiple signals at different wavelengths on an optical fiber.
Better service quality
The increase in speed of the routers mentioned above is still only to adapt to the rapid increase in data traffic. The more essential and more profound changes in the development trend of routers are: IP-based packet-switched data will rapidly replace the circuit-switched communication method that has been developed for nearly a hundred years in the next few years and become the mainstream of the communication business model. This means that IP routers must not only provide faster speeds to adapt to the rapid increase in traditional computer data traffic, but also IP routers will gradually provide the various services provided by the original telecommunications network. However, traditional IP routers do not care about or know the service type of IP packets. Generally, they only forward data packets on a first-in-first-out principle. Data of various service types such as voice calls, real-time video, Internet browsing, etc. are all treated indiscriminately. treat.
It can be seen that if IP routers want to provide all services including telecommunications and broadcasting, improving the quality of service (QoS) is the key. This is also the direction that major network equipment manufacturers (including Cisco, 3Com, Nortel, etc.) are working hard to advance. The high, medium, and low-end routers newly launched by major manufacturers all support QoS to varying degrees. For example, Cisco’s top-end 12000 series has strong support for QoS in terms of hardware and software protocols, while its newly launched low-end routers support QoS. The end product 2600 series also supports new business applications such as voice calls. In fact, QoS is not only a development trend of routers, but the entire IP network with routers as the core is developing in this direction.
The concept of "three nets in one" is the product of this direction. However, networks with traditional IP routers as the core can no longer adapt to the trend of "three networks in one". The developed countries led by the United States are advancing the research and development of network technologies that can provide better and faster service quality. The research and development of routers is the key, and the company has become the main driving force to promote this technology.
The support for QoS comes from both software and hardware. In terms of hardware, faster forwarding speed and wider bandwidth are the basic premise. In terms of software protocols, after recent efforts, the following results have been produced:
①IPv4 header service type field
There is a 3-bit area in the IPv4 packet header to identify the priority of this IP packet. Based on this priority, the IP router can determine the forwarding priority order of different IP packets. It can be said that since the date the IP protocol was formulated, a mechanism has been reserved for providing better QoS in the future.
②RSVP (Resource Reservation Protocol) and the corresponding series of agreements
This is a profound step forward for IP routers to provide better service quality. Traditionally, IP routers are only responsible for packet forwarding and know the addresses of neighboring routers through routing protocols. RSVP is similar to the signaling protocol of a circuit-switched system. It informs each node (IP router) that a data flow passes through, and negotiates with the endpoint to provide quality assurance for this data flow.
③DiffServ (DifferenTIated Service)
For the above two reasons, the IETF has introduced another QoS strategy-DiffServ. At present, the framework of DiffServ has been basically determined, and Internet2 in the United States has also selected DiffServ as its QoS strategy. Compared with DiffServ, RSVP is an Integrated Service (centralized control strategy), while DiffServ is a decentralized control strategy, the essence of which is to control only each behavior in the path. The terminal application equipment negotiates with the edge router through the SLA (Service Level Agreement) to obtain the guaranteed service level for its application data flow.
④Multi-protocol label exchange
MPLS is also used to solve the QoS problem, but its coverage is the core network router. In order to establish a reasonable exchange path between core routes, core routers need to exchange traffic and other status information regularly.
Smarter management
With the explosive growth of network traffic, increasing network scale, and increasing requirements for network service quality, network management systems on routers have become increasingly important, and network connections have become an indispensable part of daily work and life. . To maximize the use of bandwidth while ensuring quality, early detection and diagnosis of equipment failures, and quick and easy configuration changes as needed, these network management functions have increasingly become important factors that directly affect the interests of network users and network operators. In the seven-layer network protocol model, network management belongs to high-level applications. At present, an important development trend of network management of various manufacturers is to develop toward intelligence. And intelligence is reflected in two aspects, one is the intelligence of information interaction between network devices (routers); the other is the intelligence of information interaction between network devices and network managers. In the general trend of network management intelligence, the two technical concepts of "policy-based management" and "traffic engineering" are currently the most eye-catching. All router manufacturers in their newly launched products all boast that their network management supporting systems have or partially have these two functions.
