Technical Problems Facing the Development of Ultra HD Video Surveillance
In the field of security, factory-installed ultra-high-definition video surveillance has very promising application prospects, but only after solving several problems that hinder application transmission, computing power, algorithms, storage, security, etc., the application prospects will become Be clear. On the other hand, the traditional network is undergoing earth-shaking changes, with unprecedented improvements in computing and storage capabilities. Algorithms are further hardware-based and intelligent. Security issues have never become the will of the country like today. All these are breakthroughs in technology for ultra-high-definition video surveillance. Here comes a bright future and a drive for progress.
The first problem facing the installation of ultra-high-definition video surveillance in the villa was the transmission problem. Due to the large resolution of 4K video, for a frame rate of 25fps, the code rate is about 4 times that of high-definition video (1080P) under the same encoding specifications, and the transmission requirements have also been increased several times. Even with more advanced encoding methods such as H.265, the transmission volume of UHD video cannot be underestimated due to improvements in color depth, frame rate, and resolution. In the era of 8K ultra-high-definition video, its transmission volume will increase exponentially. Therefore, increasing the bandwidth, that is, increasing the end-to-side throughput and increasing the transmission capacity of the intermediate link are the primary issues facing ultra-high-definition video surveillance .
(1) Increase end-to-side throughput
The end is the receiving end and the sending end of the ultra-high-definition video. It is critical to increase the uplink and downlink capabilities of the network card at both ends. The uplink and downlink capabilities are constrained by the following factors: network card performance, buffer size and scheduling mechanism, network protocol stack working efficiency, the throughput capacity of the ultra-high-definition video monitoring application process itself, and video receiving and sending strategies.
① Network card performance optimization
School installation monitoring In order to ensure the quality of surveillance video transmission, we use a single Gigabit card with an effective uplink and downlink transmission rate of 60%. In the case of a single Gigabit card, for 4K UHD video encoded by H.264MainProfile, even if its bit rate is only 4 times that of 1080P, it will be close to 30Mbps. Therefore, a single Gigabit network card can only carry 20 channels of 4K Ultra HD video . This may not be a problem for browsing clients, but it is not enough for streaming media servers. Therefore, it is particularly important to upgrade from a Gigabit card to a 10 Gigabit card, or to bundle multiple Gigabit cards to extend the uplink and downlink capabilities.
On the other hand, warehouse installation monitoring can completely “offload” functions such as network packet soft check, encryption and decryption, DPI and other functions performed by software. This is our familiar hardware offload acceleration technology. These functions are designed and described in hardware language through the SOC, and it is a wise move to implement ASIC circuits in the SOC.
② buffer optimization
Chongqing factory installation monitoring For the network transmission application of video surveillance, streaming media servers account for the bulk of the traffic. Therefore, it is increasingly necessary to improve the mechanism and performance of streaming media services. As an intermediate medium for the network card to interact with the operating system and application software, the buffer area should be improved accordingly.
a.HugePage mechanism: The allocation granularity of memory pages in the operating system is 4KB, which is obviously not enough for ultra high-definition video, so the large memory page mechanism or even the giant page mechanism is selectively enabled to make the allocation granularity reach several MB or even 1GB to reduce the system overhead caused by memory page switching, which is of great significance to both the sender and receiver.
b. DMA mechanism: DMA is the direct memory access mechanism. DMA can abandon the traditional "network card cache-> main memory-> CPU cache" transmission path, and instead use the DMA controller to establish a mapping between the network card cache and the CPU three-level cache to achieve fast data exchange. Because the slower step of main memory read and write is bypassed and the IO of the PCI-E bus is omitted twice, the read and write speed will be greatly accelerated.
③ Network protocol stack optimization
The traditional network protocol stack exists in the operating system in a kernel-driven manner, and its key working mechanisms are interrupt response, delayed process processing, and general packet processing.
Interrupt response: The traditional network protocol stack driver is based on the interrupt mechanism of the network card. The arrival and transmission of network packets are notified to the upper network protocol stack by the interrupt mechanism so that the protocol stack driver can continue to receive and send.
Delayed process processing: After the protocol stack driver responds to the interrupt, it does not include the packet receiving or sending process in the interrupt processing routine to occupy the interrupt time because the interrupt has a higher priority. If the interrupt takes too long, it will affect other priorities High-level thread execution, so the interrupt processing routine puts specific transaction tasks such as charge / send in the DPC (Delayed Procedure Call) queue, and processes it only when the interrupt priority drops, which reduces the time occupied by interrupt interruption.
Generic packet mechanism: The network protocol stack is aimed at general-purpose network packet processing. Therefore, each layer of the OSI model will be processed and verified accordingly, which is more suitable for situations where the traffic is not large and the packet types are different. On the high-definition video streaming server, the traffic is large, and the signaling packets and video packets are generally transmitted. The protocol format and encapsulation method are fixed.
The above mechanism reduces the processing efficiency of the protocol stack to a certain extent. For ultra-high-definition video streaming media servers, an improved network protocol stack can be used to bypass traditional protocol stacks, such as customizing a dedicated protocol stack driver specifically for streaming media transmission, or a protocol stack driver that grafts high-speed transmission equipment. The DPDK (Data Plane Development Kit) framework is a good choice. DPDK is a network packet processing framework based on the Intel X86 / X64 platform. It is also a set of packet bypass processing schemes. It has a high IO processing speed. It is mostly used as a forwarding driver framework for SDN high-speed switches and routers. Features and mechanisms:
a. UIO mechanism: The UIO (UserspaceI / O) mechanism runs a small part of the driver in kernel mode space (hard interrupts can only be processed in kernel mode space), and most of them run in user mode space to implement the bypass mechanism.
b. SIMD mechanism: The DPDK framework uses batch mode to process multiple network data packets simultaneously. Based on vector programming, all network data packets are processed in one cycle, which increases the processing throughput.
c. Cache optimization mechanism: Adopt Cacheline alignment, Cache data prefetching and other strategies to speed up the reading and processing of data in the cache.
d. PDM mechanism: The PDM (PoolModeDriver) mechanism discards the interrupt mode, and instead receives packets based on the interrupt + polling method, avoiding interrupt overhead.
e. Lock-free circular queue mechanism: Supports single producer enqueuing, single consumer dequeuing, multi-producer enqueuing, and multi-consumer dequeuing operations, so it can improve transmission efficiency and ensure data synchronization.
f. Processor affinity mechanism: The processor affinity (CPUAffinity) mechanism is used to bind IO threads to several CPU cores to reduce thread scheduling and switching and thus reduce switching overhead. At the same time, threads are bound to On a fixed CPU core, the hit rate of the CPU cache is greatly improved.
g. Multi-queue mechanism: Through the support of the multi-queue network card driver, each queue is bound to a different CPU core to meet the high-throughput requirements of the network card.
h. DDIO mechanism: DDIO (DataDirectIO) is a technology proposed by Intel that allows network cards and CPUs to directly exchange network data through LLC (lastlevel cache), thereby bypassing main memory, shortening the interaction process and increasing the speed of interaction. This technology is similar to the DMA mechanism, but has higher efficiency than DMA.
i. Hardware acceleration mechanism: "offload" basic repetitive soft transactions (such as computational analysis tasks, TCP group tasks, and TCP segmentation tasks) to the hardware to speed up processing.
- Explanation of four common problems in network monitoring engineering
- How to install monitoring in the elevator? 1/15/2020
- Intelligent weak current engineering video surveillance system monitoring pole knowledge summary 2020/1/8 1
- Technical Problems Facing the Development of Ultra-high-definition Video Surveillance 2020/1/8 1
- Hikvision user-defined ball machine bracket thread interface parameters 2020/1/2 1
- Explanation of Chongqing monitoring system installation specifications 2020/1/2 1
- User-defined ball machine bracket thread interface parameters
- How to check the monitoring screen when the mobile phone cannot connect to the Internet? 2019/12/25
- Explanation of various alarm functions of Haikang video recorder
- Hikvision monitoring system prompts "code stream type is not supported" how to solve? 2019/12/18
- Hikvision video surveillance storage server configuration steps
- How should the network video surveillance project realize the construction of the network environment?
- Hikvision DS-4000HC, HC +, HCS, HF, MD series boards frequently asked questions2019 / 12/4
- Third-party domain name access to Hikvision surveillance cameras on the Internet
- Internet Video Surveillance Solution Based on Dynamic IP
- "The fusion of material and information, AI empowerment" as the theme of the wonderful appearance
- Top 10 most popular surveillance camera brands in 2019
- Home remote monitoring camera installation method
- What should I pay attention to when installing a video surveillance camera? 2019/11/13
- Explanation of methods for selecting security monitoring equipment