Networks tend to be integrated with each other as the network technologies are developing. A router that incorporates such features as routing, VOIP, transmission, security and terminal access is no longer something new. The routing and switching in one machine, a network device that has functions of both router and switch, comes into being with a trend towards IP access. It can effectively reduce network complexity, learning costs of IT staff, and equipment acquisition and maintenance costs since it is easy to deploy and configure and it incorporates multiple services into one platform. It gradually becomes a hot product for branches to access LANs and a mainstream product for bank branch access.

The routing and switching in one machine is a technical innovation. Currently, the mainstream products have two different architectures: integrated switching architecture and distributed switching architecture.

Integrated switching architecture: The switching part and the routing part in the routing and switching in one machine share a CPU. The switching chip is directly integrated into the router to provide common startup and configuration files.
Distributed switching architecture: The switching part of the routing and switching in one machine has a separate CPU and a separate startup file. The switching part is implemented by using an independent and mature switch platform.

The hardware principle diagrams for the two switching architectures are shown below. In nature, both of them use a switching chip to implement their switching parts, and the switching chip is connected to the primary CPU of the router through data channels. The switching part of the integrated switching architecture is managed and controlled by using the operating system of the router, while the switching part of the distributed switching architecture has a CPU, so its operating system and CPU are independent of each other.

Analysis of the principle shows that the switching part and the routing part of the integrated switching architecture have common operating system and CPU, so it is easy to manage and maintain. The hardware design cost can be reduced because the switching part has no CPU. However, it's difficult to expand its functions and it has limited functions.

The distributed switching architecture has its own operating system and CPU, and the routing part and the switching part have separate operating systems. The routing part and the switching part can be configured separately. The software is highly mature and the functions can be expanded easily. The architecture can provide all the functions of router and switch and has higher performance than the integrated switching architecture.

Compared with the centralized switching architecture, the distributed architecture has the following advantages:

1. Diversified software functions: The routing and switching in one machine is a brand-new product innovation. In the early days when such products emerge, only some simple functions such as switching port and VLAN are often used. As the services are increasing, the requirements for the switching functions will be further expanded and it is very likely to use other functions in the future. The centralized architecture provides limited functions, while the distributed switching architecture provides all the functions of the router and the switch and is easy to expand.

2. Higher stability and software maturity: Switching and routing are two network architectures independent of each other and focus on different aspects. In the integrated switching architecture, addition of any new function means innovation and requires addition of a brand-new switching function on the routing platform. Both routing and switching are originally large software architectures. Such innovation requires us to integrate the switching functions into the routing platform, resulting in a larger software system and affecting software maturity, stability and reliability. Inevitably, such architecture has limited switching functions and cannot provide complete switching functions. However, the distributed architecture has independent CPUs and operating systems at both its routing part and switching part. It has extremely high software maturity and its functions can be expanded easily. It has plenty of functions and can offer all the functions of the switch and the router.

3. More powerful performance and network attack prevention capability: The integrated architecture has only one CPU for routing. The CPU has to process both switching and routing functions, which has an influence on the performance. However, the distributed architecture has both routing and switching parts that have separate CPUs to perform appropriate functions. The switching CPU can process all the switching packets, which considerably reduces the load on the primary CPU and improves the overall performance of the equipment. Attacks on the switching packets in layer 2 are handled by the switching CPU directly, instead of being sent to the routing CPU. This dramatically enhances the network attack prevention capability of the equipment.

4. Easier to maintain: Since the distributed architecture has lots of software functions, it provides complete switching functions by default. It provides the functions that may be needed in the future in advance. This way, there is no need to frequently upgrade the software or replace devices even though the functions need to be expanded in the future. This cuts the maintenance costs.

5. Better scalability: Due to its modular design, the routing and switching in one machine with the distributed architecture can be upgraded and expanded with functions flexibly. Meanwhile, the functions can be expanded conveniently since the switching part and the routing part each have an independent operating system. However, the device with the integrated architecture tends to provide a fixed switching port. Therefore, the switching chip cannot be replaced after it is determined. The overall system has to be replaced in the event of upgrade, which makes expansion and upgrade difficult. Moreover, addition of any function may require plenty of development and test resources and is more difficult for vendors. The distributed architecture features better scalability.

In terms of configuration management, the integrated switching architecture allows the user to manage configurations in the unified CLI, which is convenient; however, because the switching part and the routing part of the distributed architecture each have independent CPU and operating system, the configurations often have to be managed separately. Users who need to configure devices frequently may feel inconvenient. Ruijie Networks has provided an excellent solution to this problem.

Ruijie Networks took the lead in launching the first distributed routing and switching in one machine R2700 in China in 2006. Later in 2007, it launched the new generation of distributed routing and switching in one machines, the RSR20 series routers, which are based on the RGOS platform. RSR20 is of modular design. In addition to the above-mentioned features, the latest released RGOS version solves the management problems, which allows the user to manage configurations for the routing and switching functional modules in the unified CLI. It has the advantage of unified management and maintenance as well as all the advantages of the distributed architecture.

The RSR20 series distributed routing and switching in one machines are available in 4 models: RSR20-04, RSR20-14, RSR20-18 and RSR20-