Today, the Internet of Things (IoT) has impacted all industries and is expected to become a $1.7 trillion market by 2020. The IoT space is based on cloud computing and data acquisition sensor networks built with mobile, virtual, and instant connectivity. Industry experts believe that it will make everything in our lives more "intelligent." IoT has penetrated into all industries: from factory automation to on-demand entertainment and wearable devices. But in most cases, this huge smart device interconnect system has not fully realized its full potential in changing the way we work.
IoT is undoubtedly a new driving force for the development of the semiconductor industry and embedded systems. Its birth has boosted the market's demand for many new enabling technologies, including:
· A new generation of ultra-low power ICs
· New wireless communication protocol
· Advanced Data Processing Technology for Analysis and Cloud Computing
As the chip moves toward smaller-scale process nodes, a semiconductor market segment that was relatively unobtrusive and now more prominent is memory. The Internet of Things and its byte-level data traffic are driving the market's demand for high-performance, low-power, ultra-small packaged memory. Another constraint imposed by IoT on semiconductors, especially memory, is security and reliability requirements. A lot of private information will be stored on wearable devices, servers and other IoT nodes.
In the past decade, the memory sector has been divided into two distinct product families: fast and low-power memory, each with its own characteristics, applications, and pricing. As long as they are willing to sacrifice power consumption or even size, OEMs can find high-speed performance memory products. For volatile and non-volatile memory that require low power consumption, and vice versa.
However, IoT changed the market's memory requirements. The demand is now high-performance, low-power devices. These devices are required to be able to perform complex operations using portable power supplies. They also need to minimize the number of pins and overall dimensions. With built-in deep power-down, deep sleep and other low-power modes, while providing a generation of performance over the next generation (ie clock frequency and feature set), the microcontroller can meet these requirements. In order to keep pace with the microcontroller, the memory must not let designers worry about trade-offs between performance and power consumption.
This article will focus on the development trend of memory in the field of retail shopping that has been affected by IoT. With the help of IoT to bring convenience to consumers, this $2 trillion market has great potential. Retail is one of the most competitive industries in the world. Millions of retailers compete for a mature customer base, so the profit margin is very low. Large shopping malls have begun to use the Internet of Things to attract customers and provide them with a personalized shopping experience. Retailers are consolidating all the equipment in the mall and corporate headquarters cloud resources. The ultimate goal is an online mall that can use the data it collects to promote, build customer loyalty, manage inventory, and increase operational efficiency.
Today's consumers are using the Internet extensively to influence their shopping decisions: from research products to online shopping to commenting on products. In the use of the Internet for shopping, retailers have lagged behind consumers. To keep pace with consumers, retailers are associating the physical and online aspects of retail so that each interaction is rewarded in order to make their shopping malls more "intelligent."
Smart POS terminal
One of the significant effects that IoT has brought to the retail sector is the smart point-of-sale (POS) terminals. In a sense, a POS terminal is a retailer using a central node of an IoT. Many leading "smart" stores use POS data to understand customer shopping trends, track inventory in real time, and help online shoppers accurately determine the product's local inventory. They also help retailers provide customized advice based on the frequency with which customers purchase specific items.
In order to track the shopper's shopping statistics, the smart POS terminal needs to connect the scanner. This means that smart POS terminals must handle several times the data processed by traditional POS terminals. Many of the latest models of smart POS terminals use the latest ARM processors that reach the Ghz level. At the same time, these terminals are mostly battery-powered portable devices. In other words, these systems need to use as little electricity as possible. In addition, since the transmitted data is highly personalized data, the highest level of data integrity is required, ie, a stricter encryption standard than the conventional terminal is required. Finally, use standard fail-safe technologies (such as lockout mode) that all POS terminals use.
POS terminals employ several types of memory: flash memory for non-volatile data storage, DRAM for cache, and SRAM for microcontroller memory expansion and battery backup configuration data logging. Sometimes even use an external MMC. Figure 1 shows a block diagram of a typical POS terminal design. In order to meet the requirements of smart POS terminals, the memory should provide the highest reliability and enough bandwidth. Not only that, in order to meet the portable requirements, the memory must also have the characteristics of low power consumption, small size.
In the past, memory development has been trying to combine fast access speeds, low power consumption, and small size features. However, with the advent of a new generation of low-pin-count interfaces such as Octi-SPI and HyperBus, there is now a bandwidth comparable to or even faster than that of fast-access memories, while competing against the power consumption of low-power memories and using the lowest Microcontroller pin memory. Another innovative technique from microcontrollers to memories such as SRAM is the introduction of deep sleep mode. For example, Cypress's PowerSnoozeTM SRAM is a Fast SRAM that has deep sleep efficiency comparable to Micropower SRAM.
figure 1. This is a block diagram of a modern POS terminal. The memory used includes flash memory, SRAM, DRAM, and SD/MMC slots.
Let's compare the power consumption and access time of two commonly used SRAM-Fast and Micropower, and Fast SRAM with deep sleep mode.
By combining fast access and deep sleep features, these memories can match the speed of SRAM and the energy efficiency of low-power SRAM. The advantages of this combination are even more pronounced in applications where SRAM is in standby most of the time.
In a typical POS terminal that uses SRAM to record configuration data, the SRAM runs for only 20% of the total operating time. If this SRAM is a Fast SRAM with an operating voltage of 3.3V, it will consume 120 watts (WH) of power when in operation and 80 WH of power in standby, with a total energy consumption of 200 WH. If it is a Fast SRAM with deep sleep mode, 120 WH of power is still consumed during operation, but the power consumption is reduced to 0.06 WH during standby, so the total power consumption is about 121 WH. In this specific example, the deep sleep option reduced energy consumption by 40%.
For a 240mAH onboard button cell battery, a standby 16Mb Fast SRAM will allow the battery to run for more than 12 hours, while a standby low power SRAM will allow the battery to run for more than 3 years, but after The limitation is that the storage speed is slower. At this point, a Fast SRAM with deep sleep mode has a significant advantage over low-power SRAMs, with more than four times the bandwidth (ie 10 ns access time vs. 45 ns access time) and no power penalty. However, regardless of the MCU or SRAM, one factor to consider when using deep sleep mode is the time to enter and exit deep sleep mode. If the time between two duty cycles is too short compared to the time it takes for the SRAM to enter or exit deep sleep mode, this method will be useless. For example, for Cypress's Fast SRAM with deep sleep mode, this interval is 300 μs (maximum). This may be the biggest obstacle to the promotion of Fast SRAM with deep sleep mode.
Another interesting trend in the memory space is that as flash becomes faster and faster, the need for cache is changing. Many micro-controller processes that require RAM can now be implemented on flash memory using XIP (Execute In Place). This means that more RAM is being used for extended memory or battery system backups. At the same time, SRAMs that have been used in both applications are increasing capacity choices. In other words, the traditionally preferred DRAM is becoming less and less important because just as the larger, faster and lower power flash memory can meet the needs of large-scale storage, the capacity is higher and the power consumption is higher. Low, smaller SRAM can also meet the needs of small storage.
Other components
There are many other components for building a smart shopping experience: various types of sensors, electronic shelf labels and beacons, storage devices, and data processing terminals for processing collected data. It is very difficult to discuss the application and memory requirements of all these devices in one article. I plan to explore the memory requirements of these devices in the near future. However, the basic requirements remain unchanged, namely low power consumption, high speed, small size, and high reliability.
Gaming Earbuds With Mic,Metal Body Earphones,Sport Metal Headphone,Wired Metal Earphone
Guangzhou YISON Electron Technology Co., Limited , https://www.yisonearphone.com