When implementing Internet of Things tech, remember: One size does not fit all
May 01, 2013
As the Internet of Things (IoT) continues to accelerate the need for cloud connected embedded devices, IoT use cases will all entail processing, sensi...
ECD: What new microcontroller technologies are available to meet the growing demand for cloud connected embedded devices?
KARIMI: The focus on the Internet of Things (IoT) is driving the growing demand for cloud connected embedded devices. Embedded processing is at the heart of the IoT. However, there are layers of embedded processing that will be happening all the way from the sensing nodes on the edge of the network, through the cloud processing. Requirements common to all types of IoT-related use cases include sensing, processing, wired or wireless communication, software, and security – most of which involve microcontrollers or microprocessors in some fashion. Use cases vary significantly, as one size will not fit all.
Specific technologies include extremely low power 32-bit MCUs with a CoreMark score of 15.9 CM/mA, which means power consumption running CM is 6.684 mA. Such MCUs perform at almost <40μA/MHz. Additionally, other technologies and strategies suited for IoT-related embedded processing include cost effectiveness for end sensor nodes and advancements in larger memory integration (both RAM and ROM), very low-power NVM technologies for read-intensive applications, seamless integration of HMI and haptics technologies, wired and wireless MCUs, and using robust tools and system software as building blocks to get to market faster. All of these include an emphasis on ease of use and system solutions. A thorough and comprehensive security and tamper-protection strategy is also a must-have at all layers of embedded processing, and this is an area that is subject to a lot of innovation.
ECD: As you look ahead for the next few years, what trends do you see in the embedded design area and how will they affect your product development plans?
KARIMI: Embedded processing is being proliferated into new applications every day and the IoT trends have added new dimensions to these trends. I expect this will continue and will impact our development plans and have implications on product size, cost, power efficiency, and ease of use. Integrated precision analog, a robust set of peripherals across a wide portfolio, differentiating flash, a wide set of tools and enablement, and quality and reliability are all requirements needed for success in the embedded space. As an example, adding embedded processing to smart pills for digestible devices or adding embedded processing for smart shoes – versus the traditional MCU and MPU markets segments such as home automation or white goods such as washers, dryers, and so on – will require changes to how we build and package our devices to meet the constraints of the applications.
Wireless MCUs, for the most part, need a 32-bit architecture, yet the expectation is that the cost, size, and power specs will be less than the 8-bit products of a couple of years ago. Add the need for an ecosystem of software providers and you can see why ARM is becoming the architecture of choice for MCU designers.
ECD: With ubiquitous connectivity dominating embedded designs, what security precautions are available to prevent unauthorized access?
KARIMI: Security technology must be coordinated across the continuum of IoT embedded processing layers. For us, this means a system-level approach to security from the sensing nodes at the edge of the network, through various layers of embedded processing, all the way through the data centers and cloud processing. It also encompasses what is needed for a safe and secure automobile. Specifically, this system-level approach requires a wide range of peripherals and hardware and software technologies to support a secure design and protect against various forms of attack, depending on the use cases as well as the security needs of that segment application. There’s no such a thing as “one size fits all” here, and some system-level techniques include use of crypto engines and secure memories that can automatically be erased upon the detection of tamper; other techniques include embedded flash locking, debug security, and boot-up checks that ensure application software integrity. Use of security protocol accelerators and processors that integrate seamlessly with ARM TrustZone – which separates and protects secured keys and environments from the application space – is also becoming popular.
On the hardware front, the addition of cryptographic engines, embedded security modules, and protocol accelerators is important. On the software front, symmetric and asymmetric crypto middleware to support the hardware cores and architectures are required. Then for the application software, HDCP and DTCP-IP are examples of the software needed for DRM and content protection. Finally, wireless MCUs and MPUs and the implementation of networking security protocols such as IPsec, SSL/TLS/DTLS, SRTP, MACsec, and so on are a must-have for that part of the portfolio, above and beyond what we just talked about.
ECD: Software development is a huge portion of each new embedded development project. What development tools and libraries do you offer developers?
KARIMI: Today software is at least as important, if not more so, as the hardware it commands, and over time, we believe software will become a key differentiator when our customers choose our devices. As a semiconductor company, it’s been our experience that it’s important to not only offer your own software tools but also leverage the third-party ecosystem built around various cores, such as the ubiquitous ARM cores.
Besides development tools there are RTOS solutions, DSP libraries, encryption libraries, connectivity libraries, and Processor Expert software generation tools. We have our own development platforms – the Tower system and Freedom Development Platform.
We also have application-specific code for a variety of end segments. For example, our wireless MCUs not only support the protocol stack and security software needed, but for select segments we support application layer profile software.
ECD: Does your company offer any educational events or online classes to help embedded designers get started with microcontroller-centric projects?
KARIMI: Yes. There are several hours of webinars and presentation materials on our website: www.freescale.com. In addition, we have domestic and regional “Designing with Freescale” events that focus on topics that are important to the embedded designer of particular geographies and regions and with specific sets of applications in mind. Lastly, on a biannual basis, we have our Freescale Technology Forum, which brings thousands of customers to meet with the Freescale engineering community and attend comprehensive classes of their choice.