What is multilayer switching technology

Multi-Layer Exchange

When I first heard the word third exchange, there was some confusion, followed by the fourth-tier exchange, the seventh-tier exchange concepts are more people to consider. In fact strictly speaking, the exchange means that the source and destination address between the connection, in the second tier of any technology can not be said to be exchange technology. The term load-balancing has largely replaced the fourth-tier exchange, just as the use of the word cognitive has largely replaced the seventh-tier exchange. However, I am afraid that the third-tier exchange term will always be so called.

Of course, the point is that the key is to recognize the benefits of these technologies in improving the performance of the network, so this article still follows the "first layer of exchange" this term.

Third-tier exchange

The third layer switching technology is also called IP switching technology, high-speed routing technology and so on. This is a mechanism that utilizes information from the third layer of Protocol to enhance the second-tier switching function. The vast majority of today's enterprise network has become the implementation of the TCP/IP protocol Web technology intranet, the user's data often over the local network across the internet transmission, so the routers often overwhelmed.

One way to do this is to install a more powerful super router, but it's too expensive, and if it's a switched network, that investment is obviously unreasonable. The goal of the third-tier exchange is that there is no need to forward packets via routers as long as there is a more direct second tier between the source and destination addresses. The third tier Exchange uses a third-tier routing protocol to determine the routing path, which can be used only once or stored for later use. The packet then passes through a virtual circuit around the router to send it quickly.

At present, the main third-tier exchange technologies are:

Ipsilon IP Exchange: IP switching technology initiated by the Ipsilon company, that is, the identification of packet flow, as far as possible in the second layer of exchange to bypass the router, improve network performance. Ipsilon improves the ATM switch, deletes the software from the controller, plus an IP switching controller, which communicates with the ATM switch. The technology is applicable to the intranet and campus network within the organization.

Cisco Label switching: Labels the packets, which are read out at the switching node to determine the packet routing path. This technology is available for large networks and the Internet.

3Com Fast IP: Focus on data policy management, priority principles, and quality of service. The Fast IP protocol ensures that the required bandwidth is available for real-time audio or video data streams. Fast IP supports other protocols, such as IPX, that can be run in a swap environment other than ATM. The client needs to have software that sets the priority level.

IBM ARIS (Aggregate route-based IP Switching): Similar to Cisco's label switching technology, the packet is attached with tags to pass through the switching network. Aris is commonly used in ATM networks and can be extended to other switching technologies. A boundary device is an entry into an ATM switching environment, containing a routing table with a third-tier routing map to a second-tier virtual circuit. The ATM network is allowed to send data through a virtual circuit with two or more computers at the same end, thereby reducing network traffic.

MPOA (Multiprotocol over ATM): A specification presented by the ATM forum. After the source client requests, the routing server performs routing calculations and gives the optimal transmission path. Then, an Exchange virtual circuit can be established to cross the subnet boundary without further routing.

At present, Cisco, 3Com, Nortel Network, Lucent, Cabletron, foundry and extreme and other companies have a relatively mature third-tier exchange products and modules launched. The following is an example of 3Com company's technology to illustrate the evolution of the third tier switching technology.

The first generation switch is a hybrid of discrete electronic components and the original language software framework. The function of the software is running on a processor with fixed memory, and the function of the software is increasing with the improvement of management support and protocol function. When the user's daily business is more dependent on the network, the traffic on the network increased, network equipment became a bottleneck.

Although processors and memory are becoming faster and more efficient, they are still not up to the level of increased traffic. The first step in solving the problem is simplifying the network layer: replacing routers with switches to reduce the overhead of processing packets and significantly improve transaction processing speed. 3Com introduces a dedicated integrated circuit (ASIC) dedicated to optimizing the second layer of processing, which improves performance by 10 times times and lowers the overall cost of the system.

The flexible and Intelligent routing engine (FIRE) announces the advent of the third-generation switching technology. This generation is not only built on the second generation of progress, but also provides line-speed performance for third-tier routing, multicast (multicast) and user-selectable strategies (Policy), and the performance of the second and third tiers is no longer inconsistent.

Fire is a core part of 3Com's third generation third-tier switch, an innovative, integrated internetwork architecture that provides a wide range of second and third-tier functionality while providing wire-speed performance on a variety of network interface types.

Layer Fourth Exchange

End-to-end Performance and quality of service require a careful balancing of the load on all networked devices to ensure smooth flow of data between the client and the server. The second and third tier exchange products play a good role in solving the bandwidth and capacity problems of local area networks and the Internet, but this may not be enough and requires more performance, which is where the fourth tier of exchange comes in.

Layer Fourth Exchange utilizes the information in the third and fourth layers of headers to identify the application data flow session, which includes the Tcp/user Datagram Protocol (UDP) port number, the "Syn/fin" bit that marks the start and end of the application session, and the IP source/destination address. With this information, layer fourth switches can make intelligent decisions about where to forward the session transport stream. The role of layer fourth exchange is particularly important for large enterprise data centers, Internet service providers, or content providers that use a variety of different systems to support one application. Similarly, when replicating on many servers, layer fourth swapping can be very useful.

Routers and third-tier switches do not know which package is in the previous packet when forwarding different packets. Layer fourth switching technology tracks and maintains individual sessions from beginning to the beginning. Therefore, layer fourth switches are real "session switches".

Routers make forwarding decisions based on the availability and performance of links or network nodes, while layer fourth switches make forwarding decisions based on session and application layer information. As a result of this, the user's request can be forwarded to the "best" server based on different rules. Therefore, layer fourth switching technology is the ideal mechanism for data transmission and load balancing among multiple servers.

A switch with layer fourth functionality can function as a front-end to the "Virtual IP" (VIP) connected to the server. Each server and server group that supports a single or universal application is configured with a VIP address. This VIP address is sent out and registered on the domain Name System.

When a service request is issued, layer fourth switches to identify the start of a session by determining TCP start. It then uses complex algorithms to determine the best server to process the request. Once this decision is made, the switch links the session to a specific IP address and replaces the VIP address on the server with the server's real IP address.

Each layer fourth switch holds a connection table that matches the selected server's source IP address and the source TCP port. Then layer fourth switches forward the connection request to this server. All subsequent packets are again mapped and forwarded between the client and the server until the switch discovers the session.

In the case of layer fourth exchange, access can be connected to a real server to meet user-established rules, such as having an equal number of accesses per server or allocating transport flows based on the capacity of different servers.

At present, the general single function load Balancing product can connect 400 to 800 accesses per second. A new generation of products with second-and fourth-tier functionality (hardware-based load-balancing functions using customized ASIC) are connected more than 100,000 accesses per second.

The key issue in all of this is how to determine which of the most available servers to transfer to, and how many methods are being used to make load balancing decisions. Depending on the amount of granularity required for load balancing, layer fourth switches can use a variety of methods to assign application sessions to the server. These methods include the simple weighted cycle of minimum access for weights, the measurement of round-trip time delay and the closed loop feedback of the server itself, and so on.

Closed loop feedback is the most advanced method that utilizes specific system information, such as available memory, I/O interrupts, and CPU utilization, which can be automatically obtained for adapter drives and layer fourth switches. The current closed loop feedback mechanism requires that software agents be installed on each server.

Layer fourth switches are completely different in form and function from dedicated load balancers. The traditional hardware-based load balancer is an optimized two-port device with a speed of 45Mbps.

The layer fourth switch is designed for use in high-speed intranet applications, and it supports 100Mbps or gigabit interfaces.

Layer Fourth Exchange supports other functions besides load balancing, such as the transport flow control function based on application type and user ID. Using multi-level queuing technology, layer fourth switches can label the transport flow and assign priority to the transport stream according to the application. In addition, layer fourth switches are placed directly on the front of the server to understand the application session content and user permissions, making it an ideal platform to prevent unauthorized access to the server.