Comprehensive Understanding of vro Parameters

Broadband routers are still quite common. So I have studied the various parameters of the Broadband Router. I will share them with you here and hope they will be useful to you. Throughput) indicates the amount of data that the router can process per second. For example, the router works like a post office package business. The Post Office sends parcels to the post office to send the items, senders, recipients, and other information to the post office. The post office packs the items, and paste the formatted package form, check the correctness and then deliver it to the destination. Package receiving is the opposite process.

The vro is basically the same, but what you send and receive is data. Throughput is equivalent to the package processing capability in the post office unit time, which is an intuitive reflection of the router performance. But at the same time, this data is also the most hidden name hall. First, it should be noted that the Throughput of the router must be a LAN-to-WAN Throughput. The router processing is required only when the data flows out or into the LAN to represent the performance of the router.

Instead of LAN-to-LAN, this represents the performance of small switches inside the vro, which makes no sense at all. This is like the parcel processing capability in the post office unit time. It must refer to the number of parcels that are delivered and run into the post office, rather than the number of packages that are migrated from one post office room to another.

Secondly, the router Throughput should be the test data obtained when NAT network address translation is enabled and the firewall is disabled. This is because NAT is the most basic and core function of a Broadband Router. If you do not enable NAT, it cannot be a Broadband Router. The software design directly affects NAT efficiency and router performance, therefore, Throughput enabled by NAT is meaningful. The firewall should be regarded as an advanced function attached to the Broadband Router. Some product firewall rules are complex and can filter many items, and some product rules are less and simpler. There are many and complex rules. The time that the CPU uses to filter data is long, the rules are few and simple, and the time that the data is filtered is short. This has a huge impact on the Throughput test data. To be fair, it is reasonable to disable the firewall when testing the router Throughput, especially when comparing the performance of different products. The evaluation of the firewall is generally placed in the function comparison rather than performance comparison.

There are two common Throughput Test Methods: Smartbits test and Chariot test. Smartbits testing uses the world's most authoritative network device tester Smartbits2000 or 6000, with a professional test software SmartApplication, and uses continuous UDP packets to measure data packets of different sizes, the number of packages that the router can process per second. When a router processes data packets, the main time is spent on the packet header and packet end. For data packets of different sizes, the number of packets that the router can process per second does not vary much. This is like processing parcels at the post office. It mainly takes time to process the sender, recipient, and item information and check. A lighter package will not affect the processing speed. For example, a 10000 Byte packet can process 20000 packets per second, but cannot process 10000 64 Byte packets per second. Instead, it can process only a little more than 10100 packets, for example, packets. For example, in some cases, the maximum number of 1518Byte packets processed by a vro is 8000 per second. In theory, the maximum number of 1518Byte packets processed by a vro is 8127 per second ), the Throughput value is 100 M * 8000/8127 = 98.44 M. Therefore, the Throughput value of the router is as high as 98.44 M; in fact, this router processes the smallest 64-byte packet with 11000 theoretical limit values per second (148810). The conversion is only 100 M * 11000/148810 = 7.39 M, and the difference is more than 13 times! Similarly, the router's "Throughput Up To 97M" may also be a 1518Byte package of 97 M, but the 64 byte package may only be 7.39 M!

More importantly, packet forwarding can truly reflect the Throughput capability of the router! At present, most mainstream vrouters on the market process large packages such as 512, 1024, and 1518Byte, and Throughput can be similar to the line speed. Of course, there are also some fake vro with no forwarding capability ), when testing packets such as 256, 128, and 64Byte, the performance difference is immediately reflected. In most vrouters, packet forwarding is critical in actual applications. For example, when playing games in Internet cafes, each game command is operated, and every trigger is triggered, all packets are forwarded, if the ability of routers to process packets is not strong enough, it is difficult to cope with a large number of packet forwarding needs of the game, the game is not smooth to play. Therefore, enabling the Throughput of the LAN-to-WAN of a 64-Byte packet in NAT is meaningful.

In addition, according to the current hardware widely used by broadband routers, it is basically not enough to reach the 64 byte packet Throughput line speed. Therefore, some products claim that the line speed can be reached, it may be because the so-called NAT hardware acceleration technology is adopted, and Throughput is very nice, but the cost is that all data does not pass the CPU and cannot be filtered. The firewall does not work, vro does not have any management functions. This is like the post office does not check what is in the package, whether there are inflammable and explosive items, whether the address is Mars, whether the recipient is Bin Laden, and send all the mails as per the order.

The Chariot test uses two servers to install the NETIQ software Chariot. One is connected to the WAN port of the router and the other is connected to the LAN port, calculate the Throughput of A vro by calculating the time when a script file of the predefined length and format is transmitted from one server to another without error. The knowledge here is big. First, it is related to the server performance. Dishonest manufacturers use high-performance servers to test their own products, while others use low-performance servers to test their products, deliberately devalue data. Second, the number of established connections. The more established connections, the higher the performance requirement on the router software, and the poor software writing. When the number of connections increases, the Throughput will decrease linearly, it may even lead to connection failure, while vrouters with well-written software are stable and may even rise slightly as the number of connections increases.

Third, the Chariot test is not like The Smartbits test. It is a UDP "package", but a data "stream". The data encapsulated into a TCP packet is related to the server operating system, but it is generally encapsulated as a large 1518Byte package, so the test data is much better than the data in the 64 Byte packet tested with Smartbits, which is confusing. 4. The result of the Throughput test on the train ot is the effective data load, excluding the loss of the TCP protocol, the frame interval, the response, and the loss of the train OT System. The typical loss is calculated by about 6 M based on theory, that is to say, even if you are connected to a vswitch capable of wire rate forwarding, instead of a vro, the measured Throughput may only be around 94M, which is the theoretical limit value, therefore, if it is declared that the export ot test Throughput of A vro product reaches 97 M and 98 M, it is undoubtedly a message to deceive consumers.

Advanced User:

Throughput is the Throughput of the LAN-to-WAN. Generally, it is the test data obtained when NAT is enabled and the firewall is disabled. There are two methods: Smartbits test and Chariot test, the results can be different. According to the current general hardware capabilities, The Smartbits big package test data can easily reach the line speed, just like testing the intelligence of college students. The question is 1 + 1 = ?, Everyone can answer this question. Therefore, we need to make an objective evaluation. The question should be difficult. We should compare the 64-byte packet test data to determine whether to compete. Of course, if there is NAT hardware acceleration, you need to turn it off before testing.

As for Chariot testing, it should be fair to perform the same test environment and method and the same number of connections. In actual network applications, it is almost impossible to establish a single connection, therefore, it is best to test Chariot in Multi-connection mode. Generally, you can select 100 pairs for connection. In short, a single Throughput data is meaningless. It must be noted in what method the data was tested to be useful, the comparison of Mutual performance must be fair and meaningful only in the same test environment and method. The best way is to test the Smartbits test to enable the LAN-to-WAN Throughput of 64 Byte packets in NAT.