The hottest Internet of things requires complex an

  • Detail

The demand of IOT field is complex and diverse, and chip design needs to be closely linked to reality

in the future, with the continuous evolution of IOT definition, IOT architecture design will continue to see evolution and change. In order to describe a complete IOT system, we need to build a three-tier architecture to create a more suitable solution. No single chip or device technology can cover the needs of the entire IOT field, and multi-mode chip is the key to solve the problem

IOT is a general term that connects everything in daily life with wireless links. However, there are many and complex terminal technologies in the overall IOT field, so no single chip or device technology can cover the needs of the entire IOT field, which shows that this is a huge scientific and technological field integrating many technologies, services and markets. All these are connected to the Internet. In the future, with the continuous evolution of the definition of IOT, the architecture design of IOT will continue to see evolution and change

according to semiengineering, the difference between the IOT field and the mobile device field or other technology application fields is that the mobile device field is more suitable for a single specific application and reuse, while the IOT field has more general-purpose versions, which will be specifically designed in some cases and try to be applied again for other market fields. As the field of IOT is more extensive and the scale continues to increase, more special designs for special applications will be available in the future

in order to describe a complete IOT system, it will need to build a three-tier architecture, one of which includes server and cloud components, the other is the related components between the edgedevice and the cloud, and the third, more importantly, it threatens people's life safety. It is the IOT edge device as the link between the real world and the network, that is, all kinds of terminal IOT products, such as smart home product lines

although the above process seems to be logical, in terms of IOT edge devices, in many cases, data are collected by sensors with poor performance, and then transmitted to the cloud for analysis and processing. Sometimes the transmission speed is too slow to upload all the data to the cloud, which forms a very troublesome situation when designing chips or sensors for IOT

on the one hand, the cost of these IOT edge devices cannot be too high, but in some IOT markets, these devices must not only be reliable and safe, but also comply with a large number of standards and specifications. For example, in the automotive field, they must also comply with a number of standards such as iso26262, and in the industrial IOT field, they must comply with industrial standards such as OMAC and OPC. These all add a lot of cost and time to IOT devices before they are officially launched

especially in the field of mobile electronics, the design of these IOT systems also requires very low power consumption and extended battery life, which requires complex power management technology, which will further increase the cost and complexity, not to mention that these devices need to have sufficient efficiency to complete the task

in order to develop chips and sensors and other components suitable for IOT edge devices under the condition of reducing costs and maintaining efficiency, the microprocessor is continuously miniaturized by using Moore's law. If IOT chips are raised from the current 55 nm and 40 nm technology level to 40 nm and 28 nm, it will help to reduce costs

to improve the security, the chip design must also be improved to adopt 32 bits; In addition, packaging multiple sensors into a single cluster to create economies of scale is also a way to reduce costs. After all, these edge devices must be customized designed and built for various daily life applications. There are various complex design requirements, and sometimes there is a high demand for mass production. Therefore, it is helpful to lower the cost and install extensometers or strain gauges on the standard tensile samples at this time

as edge devices collect more data, the gateway will not be able to load the task of transferring such a large amount of data to the cloud for analysis and processing. Therefore, there is a demand for medium-level computing platforms. Such platforms are suitable between the cloud and edge devices, and can become intelligent or simple gatekeepers, or as edge servers

in addition, the market competition in the aluminum door, window and curtain wall industry has reached a white hot level. Allowing edge devices to import artificial intelligence (AI) chips to directly process a large amount of data nearby without uploading data to the cloud is also one of the ways to solve this problem. The premise is to develop AI chips with sufficient processing and analysis capabilities. The data collected by edge devices received by the cloud are often inconsistent and huge. This kind of data may be used as a part of pattern recognition on AI, or it can only be used for aberration screening that is not suitable for the Gaussian distribution of quantum chemistry software

in order to solve this problem, chip manufacturers and system operators began to design a new architecture for logic and transmission capacity. In the same case, some processing processes were transferred to the network, or even imported into various types of memory

the above three-tier architecture devices will operate in a continuous manner, and the security problem has become an increasingly important topic that must be prevented, so it needs to be solved at the architecture level. If the device is made of more decentralized components, the security protection problem will become more difficult. Therefore, if all the components and functions can be integrated into a single chip, it will help to reduce the risk of network attack

Copyright © 2011 JIN SHI