该设备的主要功能是对接收到的电信号进行再生、整形和放大，以扩展网络传输距离。由于缺乏交换机具备的“智能记忆”能力，其数据传输采用广播方式，将数据包发送至所有相连节点。根据信号处理方式可分为无源、有源和智能三类：无源集线器仅转发信号，有源集线器可放大信号延长传输距离，智能集线器则具备网络管理与路由功能。按结构和功能可分为未管理型、堆叠式及底盘集线器，从局域网应用角度则分为单中继、多网段、端口交换式等五类。其硬件特性决定了其在早期网络架构中的基础连接作用。
The main function of this device is to regenerate, shape and amplify the received electrical signals to extend the network transmission distance. Due to the lack of the "intelligent memory" capability of the switch, its data transmission adopts broadcast mode to send data packets to all connected nodes. According to the signal processing method, it can be divided into three categories: passive, active and intelligent: passive hubs only forward signals, active hubs can amplify signals to extend the transmission distance, and intelligent hubs have network management and routing functions. According to the structure and function, it can be divided into unmanaged, stacked and chassis hubs. From the perspective of LAN application, it can be divided into five categories: single relay, multi-segment, port switching, etc. Its hardware characteristics determine its basic connection role in the early network architecture.
Please open Telegram to view this post
VIEW IN TELEGRAM
网络集线器的主要功能是对接收到的信号进行再生整形放大，以扩大网络的传输距离，同时把所有节点集中在以它为中心的节点上。它工作于OSI（开放系统互联参考模型）参考模型第一层，即“物理层”。集线器与网卡、网线等传输介质一样，属于局域网中的基础设备，采用CSMA/CD（一种检测协议）访问方式。
The main function of a network hub is to regenerate, shape and amplify the received signal to expand the transmission distance of the network, and at the same time concentrate all nodes on the node centered on it. It works at the first layer of the OSI (Open System Interconnection Reference Model) reference model, namely the "physical layer". Like transmission media such as network cards and network cables, hubs are basic devices in local area networks and use CSMA/CD (a detection protocol) access method.
网络集线器属于纯硬件网络底层设备，基本上不具有类似于交换机的“智能记忆”能力和“学习”能力。它也不具备交换机所具有的MAC地址表，所以它发送数据时都是没有针对性的，而是采用广播方式发送。也就是说当它要向某节点发送数据时，不是直接把数据发送到目的节点，而是把数据包发送到与集线器相连的所有节点。
A network hub is a pure hardware network bottom-level device, and basically does not have the "intelligent memory" and "learning" capabilities similar to switches. It also does not have the MAC address table that switches have, so it does not send data in a targeted manner, but in a broadcast manner. That is to say, when it wants to send data to a certain node, it does not send the data directly to the destination node, but sends the data packet to all nodes connected to the hub.
网络集线器按照对输入信号的处理方式上，可以分为无源HUB、有源HUB、智能HUB。无源HUB：它是最次的一种，不对信号做任何的处理，对介质的传输距离没有扩展，并且对信号有一定的影响。连接在这种HUB上的每台计算机，都能收到来自同一HUB集线器上所有其它电脑发出的信号；
Network hubs can be divided into passive HUBs, active HUBs, and intelligent HUBs according to the way they process input signals. Passive HUB: It is the lowest type, which does not process the signal in any way, does not extend the transmission distance of the medium, and has a certain impact on the signal. Each computer connected to this HUB can receive signals from all other computers on the same HUB.
Please open Telegram to view this post
VIEW IN TELEGRAM
按细分卸载
统计信息在Internet上分段传输，每个数据包1,448字节，与TCP，iP和统计信息链路层报头配对。在传统情况下，CPU将预先按分段准备统计信息，并提前准备数据文件。例如，移动某个63,536字节的特定块将需要CPU创建并上传至少31个数据文件。
如果网络适配器解决了按段卸载的问题，则CPU可以将所有统计数据片段发送到网络适配器。这种技术通常称为TCP段卸载（TSO）或大段卸载（LSO）。
统计信息在Internet上分段传输，每个数据包1,448字节，与TCP，iP和统计信息链路层报头配对。在传统情况下，CPU将预先按分段准备统计信息，并提前准备数据文件。例如，移动某个63,536字节的特定块将需要CPU创建并上传至少31个数据文件。
如果网络适配器解决了按段卸载的问题，则CPU可以将所有统计数据片段发送到网络适配器。这种技术通常称为TCP段卸载（TSO）或大段卸载（LSO）。
Offload by Segment
Statistics are transmitted over the Internet in segments of 1,448 bytes per packet, paired with TCP, IP, and statistics link layer headers. Traditionally, the CPU would prepare the statistics by segment in advance and prepare the data files in advance. For example, moving a specific block of 63,536 bytes would require the CPU to create and upload at least 31 data files.
If the network adapter addresses offload by segment, the CPU can send all statistics segments to the network adapter. This technique is often called TCP Segment Offload (TSO) or Large Segment Offload (LSO).
Statistics are transmitted over the Internet in segments of 1,448 bytes per packet, paired with TCP, IP, and statistics link layer headers. Traditionally, the CPU would prepare the statistics by segment in advance and prepare the data files in advance. For example, moving a specific block of 63,536 bytes would require the CPU to create and upload at least 31 data files.
If the network adapter addresses offload by segment, the CPU can send all statistics segments to the network adapter. This technique is often called TCP Segment Offload (TSO) or Large Segment Offload (LSO).
校验和卸载
校验和优化算法扫描程序TCP和客户统计数据报告协议数据文件以正确捕获。校对可确保网络适配器在收到数据文件时已通过认证。
活动节点每天将交换数百至数千亿个数据文件。如果服务器处理器可以将卸载计算和校验和验证的日常任务交给网络适配器，则系统软件的总体性能仍然可以得到改善。
选择应用IPv4和ipv6协议的网络适配器，以在上载和接收时卸载校验和计算。用于安全通信的网络适配器也可以用于卸载数据加密校验和计算和身份验证的工作，而无需占用服务器处理器资源。
校验和优化算法扫描程序TCP和客户统计数据报告协议数据文件以正确捕获。校对可确保网络适配器在收到数据文件时已通过认证。
活动节点每天将交换数百至数千亿个数据文件。如果服务器处理器可以将卸载计算和校验和验证的日常任务交给网络适配器，则系统软件的总体性能仍然可以得到改善。
选择应用IPv4和ipv6协议的网络适配器，以在上载和接收时卸载校验和计算。用于安全通信的网络适配器也可以用于卸载数据加密校验和计算和身份验证的工作，而无需占用服务器处理器资源。
Checksum offloading
Checksum optimization algorithms scan TCP and client statistics report protocol data files for correct capture. Proofreading ensures that the network adapter is authenticated when receiving the data file.
Active nodes will exchange hundreds of billions of data files every day. If the server processor can offload the routine tasks of calculation and checksum verification to the network adapter, the overall performance of the system software can still be improved.
Select network adapters that apply IPv4 and ipv6 protocols to offload checksum calculations when uploading and receiving. Network adapters used for secure communications can also be used to offload the work of data encryption checksum calculations and authentication without occupying server processor resources.
Checksum optimization algorithms scan TCP and client statistics report protocol data files for correct capture. Proofreading ensures that the network adapter is authenticated when receiving the data file.
Active nodes will exchange hundreds of billions of data files every day. If the server processor can offload the routine tasks of calculation and checksum verification to the network adapter, the overall performance of the system software can still be improved.
Select network adapters that apply IPv4 and ipv6 protocols to offload checksum calculations when uploading and receiving. Network adapters used for secure communications can also be used to offload the work of data encryption checksum calculations and authentication without occupying server processor resources.
TCP卸载
TCP卸载中有一个非常有益的元素，即将所有TCP堆栈迁移到硬件配置，以便可以将此CPU计算资源返回给计算机操作系统。实际上，这表示将第3层（Internet / iP）和第4层（传输/ TCP）减少为网络适配器，并具有使用电源适配器的TCP卸载模块（TOE）解决各种数据组织的能力并转移日常任务。
该项目投资了具有彻底的TOE效果的网络适配器，而切实的利益取决于何时卸载它，从而可以释放至少先前占用的CPU资源。传统的千兆网络适配器将消耗CPU解决方案容量的一半以上，并且许多服务器应用程序都需要资源。如iSCSI在应用场景中会占用网络带宽，则Internet存储必须配备TOE电源适配器。
如果不需要详细的TOE功能，则可以选择另一个网络适配器。该电源适配器可以轻松适应TOE功能的子集，例如TSO或LRO。
TCP卸载中有一个非常有益的元素，即将所有TCP堆栈迁移到硬件配置，以便可以将此CPU计算资源返回给计算机操作系统。实际上，这表示将第3层（Internet / iP）和第4层（传输/ TCP）减少为网络适配器，并具有使用电源适配器的TCP卸载模块（TOE）解决各种数据组织的能力并转移日常任务。
该项目投资了具有彻底的TOE效果的网络适配器，而切实的利益取决于何时卸载它，从而可以释放至少先前占用的CPU资源。传统的千兆网络适配器将消耗CPU解决方案容量的一半以上，并且许多服务器应用程序都需要资源。如iSCSI在应用场景中会占用网络带宽，则Internet存储必须配备TOE电源适配器。
如果不需要详细的TOE功能，则可以选择另一个网络适配器。该电源适配器可以轻松适应TOE功能的子集，例如TSO或LRO。
TCP Offload
There is a very beneficial element in TCP Offload, which is to migrate all TCP stacks to the hardware configuration so that this CPU computing resource can be returned to the computer operating system. In practice, this means reducing Layer 3 (Internet/iP) and Layer 4 (Transport/TCP) to the network adapter, with the ability to use the TCP Offload Module (TOE) of the power adapter to solve various data organization and transfer daily tasks.The project invested in a network adapter with a thorough TOE effect, and the tangible benefit depends on when it is uninstalled, which can free up at least the previously occupied CPU resources. Traditional Gigabit network adapters will consume more than half of the CPU solution capacity, and many server applications require resources. If iSCSI will occupy network bandwidth in the application scenario, Internet storage must be equipped with a TOE power adapter.
If detailed TOE functions are not required, another network adapter can be selected. This power adapter can easily adapt to a subset of TOE functions, such as TSO or LRO.
There is a very beneficial element in TCP Offload, which is to migrate all TCP stacks to the hardware configuration so that this CPU computing resource can be returned to the computer operating system. In practice, this means reducing Layer 3 (Internet/iP) and Layer 4 (Transport/TCP) to the network adapter, with the ability to use the TCP Offload Module (TOE) of the power adapter to solve various data organization and transfer daily tasks.The project invested in a network adapter with a thorough TOE effect, and the tangible benefit depends on when it is uninstalled, which can free up at least the previously occupied CPU resources. Traditional Gigabit network adapters will consume more than half of the CPU solution capacity, and many server applications require resources. If iSCSI will occupy network bandwidth in the application scenario, Internet storage must be equipped with a TOE power adapter.
If detailed TOE functions are not required, another network adapter can be selected. This power adapter can easily adapt to a subset of TOE functions, such as TSO or LRO.
接收器调整
CPU必须花费时间来重组从单独的数据文件获得的统计数据。当需要解析多个网络端口和移动应用程序数据文件时，单独的CPU将处于繁忙状态。 RSS会将数据文件一起传输到多个物理CPU（而不是内核），以便相同的物理CPU始终解析来自同一TCP连接的数据文件。 CPU始终解决相同的数据流，并且比再次接收并形成新的统计信息要容易得多，而且速度更快。
CPU必须花费时间来重组从单独的数据文件获得的统计数据。当需要解析多个网络端口和移动应用程序数据文件时，单独的CPU将处于繁忙状态。 RSS会将数据文件一起传输到多个物理CPU（而不是内核），以便相同的物理CPU始终解析来自同一TCP连接的数据文件。 CPU始终解决相同的数据流，并且比再次接收并形成新的统计信息要容易得多，而且速度更快。
Receiver Tuning
The CPU must spend time to reassemble the statistics obtained from separate data files. When multiple network ports and mobile application data files need to be parsed, the separate CPU will be busy. RSS transfers the data files together to multiple physical CPUs (not cores) so that the same physical CPU always parses the data file from the same TCP connection. The CPU always resolves the same data stream and it is much easier and faster than receiving and forming new statistics again.
The CPU must spend time to reassemble the statistics obtained from separate data files. When multiple network ports and mobile application data files need to be parsed, the separate CPU will be busy. RSS transfers the data files together to multiple physical CPUs (not cores) so that the same physical CPU always parses the data file from the same TCP connection. The CPU always resolves the same data stream and it is much easier and faster than receiving and forming new statistics again.
设备管理器没有网络适配器的5个原因和解决方法!
原因一：网卡驱动没有安装
解决方法一：重新安装对应的网卡驱动程序。
1、 右键开始菜单，选择【设备管理器】
原因一：网卡驱动没有安装
解决方法一：重新安装对应的网卡驱动程序。
1、 右键开始菜单，选择【设备管理器】