Choosing Edge Computing Boxes: 9 Key Details Across 3 Dimensions

The booming development of artificial intelligence, the deployment of IoT devices, and the advent of 5G wireless technology have led to an increasing number of emerging scenarios demanding intelligent applications with low latency, low bandwidth, localization, high security, and low cost. These include smart cities, smart finance, smart campuses, as well as data-sensitive industries like smart transportation, smart factories, and smart healthcare. By deploying edge computing devices, computing, storage, and intelligent analysis can be placed close to the data creation point, meeting the demand for rapid processing and response to large volumes of data. Moreover, data does not need to be uploaded to the cloud, which not only reduces bandwidth costs but also better protects user data security.

So, how should industries with countless long-tail scenario demands choose suitable edge devices as important carriers for edge computing AI applications?

Important Considerations for Choosing Edge Computing Boxes

The powerful computing capabilities, data processing capabilities, as well as advantages in data security, flexibility, and scalability make edge computing boxes the preferred edge devices for meeting the application needs of edge computing scenarios. Enterprises can choose suitable edge computing boxes based on the following three major dimensions and nine details according to project needs.

First and foremost—Product Core Performance

01

Based on scenario requirements, edge computing boxes are mainly applied in various fields of video surveillance scenarios. By processing and analyzing video streams in real-time, they provide more accurate decision support and efficient business operations. Therefore, computing power, algorithm types, I/O expandability, and operating temperature are key performance indicators that must be considered when selecting edge computing boxes.

Computing Power

Computing power is productivity, determining the data analysis and processing capability, algorithm execution efficiency, and video access/processing capabilities. It is important to note that more computing power is not always better; matching the scenario requirements is the optimal solution.

Algorithms

The types, quantity, and accuracy of built-in algorithms in edge computing boxes determine whether issues in actual application scenarios can be solved efficiently and cost-effectively. Therefore, it is necessary to consider whether the built-in algorithms of the selected device can meet the corresponding requirements based on the scenario needs and data characteristics.

I/O Expandability

Edge computing boxes usually need to connect to different types of external devices; therefore, a rich set of I/O interfaces can better implement scenario applications, fully utilize existing devices, and reduce deployment costs.

Operating Temperature

Edge computing boxes are often deployed in public places such as shopping malls, airports, and stations. They typically need to ensure stable operation in an environment temperature range of -20°C to 60°C; the wider the temperature design, the more favorable it is for the device’s stability.

Secondly, an important consideration—Product Usability

02

Besides the core indicators of the product, the product’s design, power consumption, and heat dissipation balance are also very important.

Power Consumption and Heat Dissipation

In terms of heat dissipation, a fanless, all-metal design is better for the product’s heat dissipation. At the same time, products with lower power consumption can reduce energy costs and lower operational maintenance costs. Therefore, selecting products with low power consumption and high heat dissipation performance is more stable and durable.

Design

The product’s design influences the convenience of use to some extent; selecting a design that is flat on all sides and without protrusions is more conducive to the deployment and maintenance of the device.

Finally, project implementation assurance—Strength of Product Service Providers

03

Project demand docking, implementation technical support, as well as product supply stability and after-sales service response speed are important guarantees for the rapid and low-cost implementation of projects. Therefore, it is especially important to choose service providers in the edge computing field with rich product research and development experience and practical project implementation experience.

Technical Support

A comprehensive development toolkit, strong system compatibility, and enterprise-specific technical support can help projects land quickly.

Supply Stability

It is advisable to choose service providers with a high degree of localization and rich industry accumulation.

After-Sales Guarantee

Focus on whether the service provider’s after-sales response speed is timely and the professionalism of after-sales service.

Yingma Technology—Focusing on Domestic Edge Computing Boxes

Yingma Technology has been deeply engaged in the edge computing field for many years, providing multi-platform, multi-level computing power and easily deployable edge computing boxes for all scenarios, with advantages in more products, platform resources, and technology within the industry.

Multi-Platform, Stable Supply: Yingma Technology cooperates with several mainstream chip platforms in China to launch a full range of domestic edge computing boxes, such as Huawei HiSilicon, Ascend, Sanneng, and Cambricon, providing a variety of options and stable supply.

Multi-Level Computing Power: Yingma Technology’s edge computing boxes cover low, medium, and high-level computing power ranging from 2T to 38T, meeting the needs for different levels of computing power in fragmented scenarios.

The following is Yingma Technology’s key launched edge computing box matrix:

01

Model: IVP09A Intelligent Workstation

Chip: HiSilicon Hi3519DV500

Computing Power: 2.5TOPS@INT8

Video Codec: Supports 4 channels of 1080p@30fps video encoding and decoding, 4K HDMI output

I/O Interfaces: Ethernet, HDMI out, WiFi, TF, etc.

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Suitable for lightweight scenarios, cost-effective, fully domestically produced

Model: IVP07A Intelligent Workstation

Processor: 8-core ARM Cortex-A55

Computing Power: 4.8TOPS@INT8

Video Codec: Supports 24 channels of 1080p@30fps video decoding, 4K@30fps video encoding

I/O Interfaces: Gigabit Ethernet, USB3.0, SIM card, SD card, etc.

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Suitable for lightweight scenarios, a single product can connect to more devices

02

Medium-Level Computing Power Edge Computing Box

Model: IVP02E Intelligent Workstation

Processor: 4-core ARM Cortex-A55

Computing Power: Dual computing cores, reaching 10.4TOPS@INT8

Video Codec: Supports 10 channels of 1080p@30fps video decoding, 4 channels of 1080p@60fps video encoding

I/O Interfaces: 4G/5G wireless communication module and mSATA SSD expansion

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Da Vinci architecture, an upgraded solution for the Hi3559A platform

Model: IVP02D Intelligent Workstation

Chip: Cambricon MLU220

Computing Power: 16TOPS@INT8

Video Codec: Supports 16 channels of 1080p@30fps video decoding, 8 channels of 1080p@30fps video encoding

I/O Interfaces: 4G/5G, mSATA SSD, GPIO, and other interface expansions

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Powerful performance, awarded “Top Ten Innovative Products in China’s Security Industry”

Model: IVP03A Intelligent Workstation

Chip: Sanneng Third Generation TPU BM1684

Computing Power: 17.6TOPS@INT8

Video Codec: Single chip supports up to 32 channels of 1080p@30fps video decoding, 2 channels of 1080p@30fps video encoding

I/O Interfaces: 4G/5G wireless communication module and mSATA SSD expansion

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Strong single-chip encoding and decoding capabilities, strong expandability

03

High-Level Computing Power Edge Computing Box

Model: IVP03X Intelligent Workstation

Chip: Sanneng Fourth Generation TPU BM1684X

Computing Power: 32TOPS@INT8, 16TFLOPS@FP16, 2TFLOPS@FP32

Video Codec: Single chip supports up to 32 channels of 1080p@30fps video decoding, 12 channels of 1080p@30fps video encoding

I/O Interfaces: M.2 (4G/5G/WiFi/mSATA(SSD)/SIM expansions

Operating Temperature: Supports stable operation in outdoor environments from -40°C to 70°C

Advantages: Super strong computing power, supports floating-point operations, compatible with Kirin systems

Model: DEP04A Intelligent Workstation

Chip: Rockchip RK3588 + Cambricon MLU220

Computing Power: Combined computing power can reach up to 38TOPS@INT8

Video Codec: Supports 32 channels of 1080p@30fps video decoding, 16 channels of 1080p@30fps video encoding

I/O Interfaces: 4G/5G wireless communication module and mSATA SSD expansion

Operating Temperature: Supports stable operation in outdoor environments from -20°C to 60°C

Advantages: Dual CPU, stronger control performance and computing capability

In addition to the above core indicators, Yingma Technology provides a self-developed zero-code migration toolchain, achieving one-click migration of algorithms from GPU to XPU without touching the customer’s model files, with algorithm adaptation completed in as little as one day. In terms of after-sales service, Yingma Technology adheres to the principle of “24-hour online, 48-hour on-site”, addressing customer concerns promptly. Yingma Technology can also customize adaptable edge computing boxes and supporting software services according to the actual application needs of different business types, empowering enterprises to quickly acquire AI capabilities.

（Source: Yingma Technology）

About Yingma Technology

Guangzhou Yingma Information Technology Co., Ltd. was established in 2006 and is an artificial intelligence enterprise dedicated to providing AIoT products and subdivided scenario solutions that collaborate across “cloud-edge-end”.

The “Deep Yuan” AI product system under Yingma has created a full-stack AI application service architecture based on multi-level computing power hardware, driven by self-training algorithms and ecological integration, supported by AI empowerment platforms and assisted by toolchains. This architecture connects the entire chain from scenario needs to algorithm, hardware integration, business platform docking, and project delivery, providing customers with products and services that are customizable in both algorithms and computing power, promoting the widespread application of AI and edge computing in subdivided scenarios.

Yingma’s AIoT products and customized services are aimed at various industries and subdivided scenarios such as smart cities, smart transportation, smart finance, smart campuses, smart emergency response, and smart parks, providing comprehensive hardware and software support and product customization capabilities.

Yingma Technology’s core philosophy of “perceiving everything, intelligent computing empowerment” fully empowers the intelligent transformation of various industries, constructing an all-encompassing intelligent world.

Click below toread the original text for more product details!！

END