abstract: a type of core is not switch, but in the core layer (network trunk) called the core switch.What's the use of core switch?What is the difference between the core and the general switch?The role of
core core switches 】 【 switches Difference between core and ordinary switch
how many PCS to use
basic core switches under 50 units do not need to use the core switch, a router.The so-called core switches for network architecture, if is a small local area network (LAN) a computer, an 8 mouth small can call it the core switch!And core switches in the network industry refers to the network management features, throughput, strong layer 2 or layer 3 switches, a network of more than 100 computers, if want to stable and high speed running, the core switches is indispensable.
the difference between core and common switch
usually connect directly to the user in a network or accessing network part called the access layer, access layer and core layer is part is called the distribution layer or between convergence layer, access layer is designed to allow the end user to connect to the network, so the access layer switches with low cost and high port density;Convergence layer switch is more than the focal point of the access layer switches, it must be able to handle all the traffic from the access layer devices, and provide the uplink to the core, so the convergence layer and access layer switch, needs higher performance, less interface and higher exchange rate.And called the network trunk core layer, core layer is the main purpose of the forward through high-speed communication, provide reliable optimization, the backbone of the transmission structure, so the core layer switches application has higher reliability, performance, and throughput.Parameters,
core switches forward rate network consists of small packets of data in
handling of each packet consumes resources.Forward rate (also called throughput) is to point to in the case of no packet loss, per unit time through a number of packets.Throughput is like intersection traffic, layer 3 switches is one of the most important parameters, marks the specific performance of switches.If the throughput is too small, will become network bottlenecks, bring negative impact on the transmission efficiency of the whole network.Switches should be able to achieve wire-speed switching, the exchange rate to achieve data transmission speed on a transmission line, so as to maximally eliminate exchange bottlenecks.For the one billion switch, if you would realize the network without blocking transmission, requirement: throughput (Mpps) = ten billion port number & times;14.88 Mpps + one billion port number & times;1.488 Mpps + one hundred million port number & times;0.1488 Mpps
if switches nominal throughput calculation value is greater than or equal to, you should be able to achieve in the layer 3 switching and line speed.Long among them, a ten billion - bit port in the package for 64 B throughput of 14.88 the theory of Mpps, a one billion - bit port in package for 64 B throughput of 1.488 the theory of Mpps, a one hundred million - bit port in package for 64 B theoretical throughput is 0.1488 Mpps.So these values is how to get?Packet forwarding line speed measure is, in fact, per unit time send 64 B by the number of packets (the package) as the benchmark.In one billion - bit Ethernet port as an example, the calculation method is as follows: 1000000000 BPS \/ 8 bit\/(64 + 8 + 12) B = 64 PPS Ethernet frames for 64 B, 8 B frame head and should be considered 12 B fixed overhead clearance of frames.Thus, wire speed one billion package Ethernet port forwarding rate is 1.488 Mpps.Ten billion line speed Ethernet port rate of packet forwarding, 10 times, for one billion Ethernet 14.88 Mpps namely;Wire speed and fast Ethernet port rate of packet forwarding, is one over ten of the one billion Ethernet, namely 0.1488 Mpps.Bandwidth is
2, lagging bandwidth
exchange interface between processors or interface card and the data bus to the maximum number of data throughput, the lanes like overpasses have combined.Due to the communication between all ports need to be done through the backplane, so back can provide bandwidth, will become the bottleneck of concurrent communication between ports.The greater the bandwidth, provided to each port of the available bandwidth, the greater the speed of data exchange is larger;Bandwidth is smaller, the available bandwidth for each port is smaller, data exchange speed will be slow.That is to say, the backplane bandwidth determines the data processing ability of switches, backplane bandwidth is higher, the stronger the ability can handle data.Backplane, therefore, the greater the bandwidth, the better, especially for those convergence layer and the center switch.If you would realize full-duplex non-blocking transmission network, must meet the requirements of minimum back bandwidth.Its computation formula is as follows: back bandwidth = port number & times;Port rate & times;2: for the layer 3 switches, only forward rate and the back all reach the minimum required bandwidth, is a qualified switches, both be short of one cannot.
3, exchange of four exchange
fourth layer to achieve rapid access to network services.In four layers exchange, decided to transfer is not only based on MAC address (the second bridge) or source\/destination address (layer 3 routing), and include TCP\/UDP port number (layer 4) application, was designed for high-speed Intranet applications.Four layers exchange in addition to the load balancing function, also support based on the application type and transmission flow control function of the user ID.In addition, the front-end layer 4 switches directly placed in the server, it understand application session content and user privileges, thus making it an ideal platform to prevent unauthorized access server.
redundant module redundancy ability is the guarantee of the safe operation of the network.Any vendor can guarantee that its products are not in the process of running down.And failure can quickly switch would depend on redundant equipment capacity.Important components for core switches shall have redundant capacity, such as management module redundancy and power supply redundancy, such ability can ensure a stable operation of network in the largest extent.
use HSRP routing redundancy, VRRP protocol to ensure that the load sharing core equipment and hot backup, in a core and double funnel switch switch failure, three layer routing device and virtual gateway can quickly switch, realizes the double line redundancy backup, to ensure the whole network stability.Core switch is the core of the entire network and the heart, if there is a fatal failure, will lead to the local network paralysis.