Vswitch technology overview and Application Analysis

Vswitch classification and functions　

A vswitch is the "cornerstone" for building a network platform, also known as a network switch. It is also a type of hub, but it is quite different from common hub functions. A general hub only serves to receive and send data, while a switch can intelligently analyze data packets and choose to send them out. For example, we have sent a batch of packets specially sent to someone. If it is in a network environment using a general hub, everyone can see this packet. In the network environment where a vswitch is used, the vswitch analyzes the packet to whom it is sent, and then packs and encrypts it. At this time, only the packet recipient can receive the packet.

In a broad sense, there are two types of switches: WAN switches and LAN switches. WAN switches are mainly used in the telecom field to provide basic communication platforms. LAN switches are used in local networks to connect terminal devices, such as PCs and network printers. The transmission media and transmission speed can be divided into Ethernet switches, Fast Ethernet switches, Gigabit Ethernet switches, FDDI switches, ATM switches, and wildcard ring switches. Large-scale applications can be divided into enterprise-level switches, department-level switches, and working group switches. Generally, enterprise-level switches are rack-mounted, and department-level switches can have fewer rack-mounted slots) or fixed configurations, the working group-level switch is relatively simple with fixed configurations ). On the other hand, from the perspective of application scale, enterprise-level switches support large enterprise applications with more than 500 information points as enterprise-level switches, and department-level switches support Switches of medium enterprises with less than 300 information points, switches that support less than 100 information points are working group-level switches. Unless otherwise stated below, all the mentioned switches refer to LAN switches.

As we all know, switches work on the data link layer, the second layer of the OSI reference model. Their main functions include physical addressing, network topology, error verification, frame sequence, and traffic control. The physical address corresponds to the network address. The network topology includes the description of the data link layer and the physical connection mode of the device, for example, a star or bus topology, an upper-layer protocol warning is triggered when a transmission error occurs during error verification, and the data frame sequence is reorganized and transmitted except the sequence; throttling can delay data transmission so that the receiving device does not crash because it receives information that exceeds its processing capability at a certain time point. At present, the switch also has some new functions, such as VLAN support, link aggregation support, and even some firewall features. This is the function of the layer-3 switch. The layer-3 Switch adds the routing function when dividing VLAN based on the protocol.

Switch Technology Status Quo and trend analysis　

Layer-3 switching is the key to adopting Intranet. It combines the advantages of layer-2 switches and layer-3 routers into a flexible solution to provide line rate performance at all levels. This integrated structure also introduces policy management attributes, which not only associate Layer 2 with Layer 3, but also provides traffic prioritization, security, and a variety of other flexible functions, such as link aggregation, dynamic deployment of VLAN and Intranet. Layer-3 switches are divided into three parts: interface layer, switch layer, and route layer.

The interface layer includes all important LAN interfaces: 10/100 M Ethernet, Gigabit Ethernet, FDDI, and ATM. The switching layer integrates multiple LAN interfaces with policy management. It also provides link aggregation, VLAN, and Tagging mechanisms. The routing layer provides the main LAN routing protocols: IP, IPX, and AppleTalk. Through policy management, the routing layer provides the traditional routing or pass-through layer-3 forwarding technology. Policy management and administrative management enable the network administrator to adjust the network according to the specific needs of the enterprise.

Compared with Layer 3, the level 2 Adoption determines the so-called network control classification. A pure Layer 2 solution is the cheapest solution, however, it provides the least control over subnet division and broadcast restrictions. Layer-3 switches provide dynamic integration support for all levels in the classification. Traditional general-purpose routers and external switches can also achieve this goal. However, compared with this solution, layer-3 switches require less configuration, less space, and less wiring, cheaper and more reliable performance.

A layer-3 Switch has all the functions of a traditional switch. The layer-3 Switch prevails. The specific technical implementation of the switch includes:

1. Programmable ASIC

ASIC is a dedicated Integrated Circuit dedicated to optimizing Layer 2 processing. It is the core of today's networking solutions and integrates multiple functions on one chip, it has the advantages of simple design, high reliability, low power consumption, higher performance and lower cost.

2. Distributed Pipeline　

With the distributed pipeline, multiple distributed forwarding engines can quickly send data packets independently. In a single pipeline, multiple ASIC chips simultaneously process multiple frames. This concurrency and pipeline improves the forwarding performance to a new level: On-Demand Unicast, Broadcast, and Multicast on all ports.

3. dynamically Scalable Memory

For advanced LAN switching products, the real performance is built on an intelligent storage system. The layer-3 Switch directly associates a portion of the memory with the forwarding engine. More interface modules are added, including their respective forwarding engines, and memory is also expanded accordingly. It also uses streamlined ASIC processing to dynamically construct the cache, which increases the memory usage. The system can also handle large burst data streams without packet loss.

4. Advanced queue mechanism　

Even if the network device has outstanding performance, it will be damaged by the congestion on the connected network segment. Traditionally, the traffic through a port must be stored in the cache of only one output queue. No matter how high its priority is, it must also be processed in the first-in-first-out mode. When the queue is full, any excess parts will be discarded. In addition, when the queue grows, the latency also increases. This feature makes it very difficult to run real-time transaction processing and multimedia applications on Traditional Ethernet. For this reason, many network device manufacturers have developed new technologies that provide different service levels on an Ethernet segment and control latency and jitter at the same time. In this way, the mechanism of queues of different levels is introduced for each port.

This queue can better distinguish different traffic levels to match the network closer to high-performance applications. Data packets such as multimedia and real-time data streams are put into high-priority queues. The Weighted Fair queuing algorithm can be used to manage high-priority queues more frequently, but it does not ignore low-priority queues. Users of traditional applications do not notice changes in response time and throughput, while those who use emergency applications receive timely responses.

5. Automatic traffic classification　

Some data streams are more important than other data streams. With Automatic traffic classification, the layer-3 switch can instruct the data packet assembly line to differentiate user-specified data streams to achieve low-latency, high-priority transmission, and avoid congestion.

6. Intelligent permission Control

The layer-3 Switch provides multiple security mechanisms and uses a traffic classifier. administrators can restrict any identified data streams, including restricting access to servers and eliminating useless protocol broadcasts. This is a breakthrough in the network technology field, that is, providing a line rate firewall.

7. Dynamic Traffic Monitoring

Traffic classification, priority processing, and resource retention enable enterprise network and Intranet administrators to focus on more important things, that is, traditional and next-generation applications. But one thing still needs to be done is traffic monitoring. Traffic monitoring is not a policy mechanism because it is actually a protection mechanism. It monitors traffic and network congestion conditions and responds dynamically to these conditions to ensure that all network elements end users and networks themselves are under control and can run best.

Many layer-3 switches use the IEEE 802.1p service level for priority processing on congested LAN. To avoid congestion, the high-performance Layer-3 Switch even uses more advanced technologies to dynamically monitor the size of the output queue, so as to detect whether a port will become congested. By controlling the queue size and congestion, the network can maintain the limit required for latency-sensitive data streams.

8. scalable RMON implementation

RMON support has become an indispensable part of active and extensive network management. The MIB defined in RFC 1757 contain statistics on the physical layer and MAC layer. The RMON 2 defined in RFC 2021 extends the collection of statistics to the network layer.

9. Vector Processing Technology

Vector processing technology is used to accelerate the processing speed of data frames. The architecture of the layer-3 Switch not only adds the control capability of the layer-3 on the layer-2, but also adds multi-faceted vector control, thus greatly enhancing the vector processing function. Layer-3 vswitches have many advantages in vector processing:

◆ Fast Frame processing speed. With ASIC-based data packet classification, forwarding, and interpretation technologies, frame decoding by software is minimized, this method can achieve much higher performance.

◆ Highly adaptive function control. Vector processing works with programmable ASIC to support future standards with minimal overhead. For example, IPv6 support is already part of Vector Logic.

◆ Enhanced management functions. Multidimensional vector processing also includes built-in Network Management Agents and RMON.

10. multi-chip CPU

In highly reliable vswitches, a dedicated high-performance RISC processor is absolutely required. In fact, the combination of frame Processing Unit (FP) and Vector Logic provides unparalleled performance.

An independent application processor AP) can assist FP. Like FP, AP is also a high-performance CPU. AP controls all operations except Frame Forwarding: High-Level bridging and routing, such as Spanning Tree and OSPF protocols, as well as SNMP and HTTP operations. The advantage of using AP and FP is obvious, because management and computing work does not affect data forwarding, thus achieving high throughput and low latency.

Through the above technical analysis, we can easily see that high performance, security, ease of use, manageability, stackability, service quality and fault tolerance are the technical features of the current switch. As video conferencing, real-time multicast, network calls, program-controlled switching, and automatic call forwarding demonstrate the emergence of a new generation of applications in the multimedia era, where should the exchange technology develop? It is certain that high bandwidth, security, service quality, and intelligence should be the technical direction pursued by the next generation switch. However, it is worth mentioning that some manufacturers are moving towards Distributed Network Computing of switches.

1. Layer-4 Switch Technology and Application