Choosing the right embedded processor can cut cost, power and increase security on a grand scale.
The smart meter is one of the best examples of a commercially viable, widely deployed Internet-of-Things (IoT) device. It highlights the main requirements of this class of product: low cost, low power, and high security. It also illustrates the main components found in all IoT devices—sensors, embedded processing, and communications capability to connect to the Internet.
The following account will examine how the right embedded processor affects these three main requirements. It will use smart meter deployment in China as a case in point, where over a recent six-year period there have been 300 million smart meters installed.
The block diagram in Figure 1 shows the hardware components common in a smart meter. In this case two analog-to-digital converters (ADCs) convert the sensed current and voltage being supplied to the homeowner. A third ADC converts the sensed analog that measures if tampering is occurring between the meter and power grid. An embedded microcontroller processes the data from the sensors and communicates the results to the power grid data center over the Internet using a serial interface. This serial communication can transmit and receive data over power line communication and/or wirelessly via WiFi, Zigbee, Bluetooth, and low power Bluetooth. The processor also controls a small display and an interface for debugging and programming.
To illustrate the impact of cost, power, and security on the smart meter, each meter can expect to operate continuously (24/7) unattended for at least 10 years. Typical chip cost for the smart meter’s application specific standard product (ASSP) is relatively low, ranging from $2 to $3 or less. Adding a board to contain the ASSP and its surrounding elements will result in board cost on the order of $5 to $10. The final cost, that of installing and if necessary servicing the meter in the field, is on the order of a few hundred dollars, primarily loaded labor for an installation truck roll. A reduction in the cost of the IP used to create the ASSP, represents a savings to the chip supplier, board vendor, and service provider.
So if a smart meter is powered by line voltage, why is power savings important? The answer lies in the cost to homeowners and the service provider over time. Each of the 300 million meters installed in China operates on 1 to 2 watts of power. Using the average kilowatt per hour price in the United States times 24 hours a day times 365 days a year, each meter in China will consume around 17.5 kilowatt hours per year to operate. Applying this average to 300 million meters operating for 10 years with an electricity cost of $0.17 cents per kilowatt per hour, that adds up to $2.98 per household in China.
While $2.98 doesn’t seem significant, when multiplied by 300 million meters, the result is $894 million in electricity cost just to run the meters. Over a 10-year time frame, this accumulates to $8.98 billion. Saving just 10% in power over 10 years results in a saving of $898 million. Using a more efficient 32-bit processor, such as the AndesCore 32-bit CPU cores, with extensive architectural power saving elements, can result in a 30% reduction in ASSP power consumption, when compared with an ASSP built around an 8051 processor core.
The other smart meter IoT requirement is security, important for the service provider as well as the homeowner. The service provider wants to prevent hackers from stealing power by finding a way to bypass the meter. The homeowner’s security concern is to ensure that the meter isn’t used to invade his privacy or interrupt the operation of the electrically controlled elements in his home. The latter will be a greater concern as the smart meter begins being used to the home to provide add-on services: home security, energy management, and other smart home functions.
CPU core architectures need other functionality, including physical security, to provide the security from hacking that the smart meter requires when it begins providing these add-on services (see figure 2). The secure microprocessor unit at the center of the S8 provides functions not found on competitive 32-bit CPU core offerings including physical security functions such as data and address scrambling and differential power analysis protection. The first defends against hacks that target the interface between CPU and memory. The second guards against hacking the program by observing the power use signature of the CPU.
Another security vulnerability is the debug port found on most chips for software upgrades and maintenance during design, and for bug fixes in the field. To eliminate the vulnerability, many SoC designs remove the debug port before final silicon. However, a secure debug in an embedded debug module will allow designers a means of fixing software in the field while guarding against physical hacking attacks.
As illustrated by the smart meter example, today’s Internet-of-Things device needs embedded processing capability that can deliver low cost, low power consumption and high security. Older 8-bit CPUs currently used in many IoT devices lacks the processing power and architectural features to efficiently deliver these three critical IoT functions. Furthermore, many of the 32-bit embedded cores designed for server and smart phone applications are lacking the architectural features to deliver low cost, hardware security, and ultra low power savings. Only the AndesCore series of high performance 32-bit CPU cores was designed in the past decade with all these IoT elements included.