{"id":3540,"date":"2022-12-19T03:56:41","date_gmt":"2022-12-19T03:56:41","guid":{"rendered":"https:\/\/www.fibermall.com\/blog\/?p=3540"},"modified":"2023-12-07T03:08:26","modified_gmt":"2023-12-07T03:08:26","slug":"what-is-rdma","status":"publish","type":"post","link":"https:\/\/www.fibermall.com\/blog\/what-is-rdma.htm","title":{"rendered":"What is RDMA?"},"content":{"rendered":"

As we all know, the explosive growth of Internet data has brought great challenges to the processing capacity of data centers.<\/p>\n

Computing, storage, and network are the three driving forces driving the development of data centers.<\/p>\n

With the development of CPU, GPU, and FPGA, computing power has been greatly improved. Storage With the introduction of solid state drive (SSD), data access latency has been greatly reduced.<\/p>\n

However, the development of the network is obviously lagging behind, the transmission delay is high, gradually becoming the bottleneck of data center performance.<\/p>\n

\"computing<\/p>\n

In a data center, 70% of the traffic is east-west traffic (traffic between servers). This traffic generally processes data flow during high-performance distributed parallel computing in data centers and is transmitted over TCP\/IP networks.<\/p>\n

If the TCP\/IP transmission rate between servers increases, the performance of the data center will also increase.<\/p>\n

\"70%<\/p>\n

TCP\/IP transfer between servers<\/b><\/strong><\/span><\/p>\n

The process for server A to send data to server B in the data center is as follows:<\/p>\n

\"TCP\/IP<\/p>\n

    \n
  1. The CPU control data is copied from A’s APP Buffer to the operating system Buffer.<\/li>\n
  2. CPU Control data Add TCP and IP headers to the operating system (OS) Buffer.<\/li>\n
  3. Add TCP and IP packet headers to send the data to NIC, and add Ethernet packet headers.<\/li>\n
  4. The packet is sent by the network adapter and transmitted to the network adapter of server B over the Ethernet network.<\/li>\n
  5. The network adapter of server B unloads the Ethernet header of the packet and transfers it to the operating system Buffer.<\/li>\n
  6. The CPU unloads TCP and IP packet headers in the operating system Buffer.<\/li>\n
  7. The CPU controls the transfer of uninstalled data to the APP Buffer.<\/li>\n<\/ol>\n

    As can be seen from the data transmission process, data is copied several times in the server’s Buffer, and TCP and IP headers need to be added or uninstalled in the operating system. These operations not only increase the data transmission delay but also consume a lot of CPU resources, which cannot meet the requirements of high-performance computing.<\/p>\n

    So, how to construct a high-performance data center network with high throughput, ultra-low latency, and low CPU overhead?<\/p>\n

    RDMA technology can do that.<\/b><\/strong><\/p>\n

    What is RDMA<\/b><\/strong><\/span><\/p>\n

    Remote Direct Memory Access (RDMA) is a new memory access technology that allows servers to read and write memory data from other servers at a high speed without time-consuming processing by the operating system \/CPU.<\/p>\n

    RDMA is not a new technology and has been widely used in high performance computing (HPC). With the development demand for high bandwidth and low delay in data centers, RDMA has been gradually applied in some scenarios that require data centers to have high performance.<\/p>\n

    For example, in 2021, the shopping festival transaction volume of a large online mall reached a new record of more than 500 billion yuan, an increase of nearly 10% compared with 2020. Behind such a huge transaction volume is massive data processing. The online mall uses RDMA technology to support a high-performance network and ensure a smooth shopping festival.<\/p>\n

    Let’s take a look at some of RDMA’s tricks for low latency.<\/p>\n

    RDMA directly transfers the server application data from the memory to the intelligent network card (INIC) (solidified RDMA protocol), and the INIC hardware completes the RDMA transmission packet encapsulation, freeing the operating system and CPU.<\/p>\n

    \"the<\/p>\n

    This gives RDMA two major advantages:<\/p>\n