Skip to Main Content

Strategic Business Insights (SBI) logo

Internet of Things November 2018 Viewpoints

Technology Analyst: Guy Garrud

RISC-V: An Arm Alternative?

By David Strachan-Olson
Strachan-Olson is a consultant with Strategic Business Insights.

Why is this topic significant?

The Arm ISA dominates the design of low-power processors, but some prominent companies are now supporting an open-source ISA: RISC-V. Currently, RISC-V is not a serious alternative to Arm but instead enables companies to develop custom architectures and processors for specific applications.

Description

A core part of a computing system is its instruction-set architecture (ISA). An ISA provides an interface between software and hardware and determines many characteristics of a processor, such as supported low-level instructions and data types. The Arm family of instruction sets—A64, A32, and T32—has become the most used ISA in the world, and Arm-based processors support a huge variety of devices, including smartphones and Internet of Things (IoT) devices.

RISC-V (five) is an open-source ISA developed by researchers from the University of California, Berkeley, and is an emerging alternative to traditional ISAs. The researchers based RISC-V on established reduced-instruction-set-computing principles. The open nature of RISC-V means that anyone can design, manufacture, and sell RISC-V chips and software. The ISA is simple but also expandable, so organizations can develop new instructions to suit their specific needs but do not have to open source their extensions.

A number of prominent companies—including Qualcomm, Google, NVIDIA, Samsung, Microsemi, NXP Semiconductors, and Western Digital—have joined the RISC-V Foundation. A number of small companies—including SiFive, GreenWaves Technologies, and Andes Technology—are developing processor cores based on RISC-V. Western Digital uses a billion processor cores in its devices annually and plans to transition all processor development to RISC-V in the coming years. NVIDIA and SiFive recently announced a partnership to integrate a SiFive RISC-V core with NVIDIA's deep-learning acceleration technology in a system on a chip to enable artificial-intelligent edge applications. GreenWaves Technologies is developing a RISC-V chip that offers multiple cores and low-power operation for machine learning on devices.

Implications

For many IoT applications, Arm-based processors have become popular with industry because of their low power consumption, processing capabilities, and the large software and hardware ecosystem that supports Arm processors. Arm does not fabricate processors itself and instead licenses its ISAs and core designs to other companies, which then tweak and customize the designs before having another company fabricate the physical processor. RISC-V allows companies to develop custom processors without the need to pay a licensing fee for the ISA. However, companies have to develop their own processor designs or license designs from other companies, such as SiFive. The creators of RISC-V claim that the ISA has many benefits over older ISAs, and they hope that a RISC-V ecosystem will develop that will give hardware engineers a wide range of products to incorporate into devices.

Impacts/Disruptions

In a recent presentation, David Patterson, vice chair of the Board for the RISC-V Foundation, claimed that a new "Golden Age for Computer Architecture" is coming to counteract the slowdown of Moore's law. Patterson envisions the development of domain-specific architectures (DSAs) that perform specific tasks very efficiently because they are not general purpose. Patterson believes that an open ISA will enable companies to develop and improve DSAs at an accelerated rate. DSAs could be very beneficial for IoT applications because companies can develop an architecture and processor tailored to a device's precise task, which could improve performance and power efficiency.

Scale of Impact

  • Low
  • Medium
  • High
The scale of impact for this topic is: High

Time of Impact

  • Now
  • 5 Years
  • 10 Years
  • 15 Years
The time of impact for this topic is: Now to 10 Years

Opportunities in the following industry areas:

Semiconductors, custom processors, low-power devices, embedded machine learning, smart environments, cloud computing

Relevant to the following Explorer Technology Areas:

Digital Twins of Spaces

By David Strachan-Olson
Strachan-Olson is a consultant with Strategic Business Insights.

Why is this topic significant?

As digital-twin platforms continue to develop, organizations are beginning to apply digital-twin concepts to connected spaces and environments to understand and improve operations.

Description

A digital twin is a virtual representation of the elements and dynamics of a device, machine, or space. Typically, the virtual representation holds information about the physical system from its initial design through its operating life—including data from sensors connected to the asset, configuration data, and maintenance records. Digital twins provide operators with a method to visualize data associated with a system and perform analysis and simulations without needing to shut down and reconfigure the physical asset.

Most digital-twin platforms focus on individual machines, but a number of companies are now offering tools to create digital twins of spaces and environments. Potential spaces include office buildings, warehouses, factories, hospitals, schools, parking lots, streets, and even entire cities. Digital twins of spaces provide organizations with a platform to organize and visualize information and often can analyze processes to improve operations. Information to populate the virtual version of the space can come from a wide variety of sources, such as sensors, external data sets, and blueprints.

Microsoft recently revealed Azure Digital Twins—a platform that supports comprehensive digital models and spatially aware solutions of physical environments. A number of third-party companies have already used the Azure Digital Twins platform to create software tools and platforms for other organizations. The software tools provide a range of smart-space functionality, including building ownership and operations tools, energy and climate-control management, space utilization, and electrical-grid operations.

Implications

Early applications of digital twins focused on expensive and complex pieces of industrial equipment, such as jet engines and wind turbines. Industrial equipment was a good match for the digital-twin concept because it has complex designs and already included a large number of sensors. However, the traditional digital-twin approach is less useful in application to small individual devices such as thermostats, motion detectors, and lights. Instead, a model of the entire environment with aggregate information from many devices and sensors is more important. Microsoft believes that modeling the complex interactions among people, places, and things can unlock new opportunities, create new efficiencies, and improve public and private spaces.

Impacts/Disruptions

Expanding the techniques of digital twins beyond individual machines to spaces was likely an inevitable evolution of this concept. As more connected devices come online, companies will need software and platforms to visualize and analyze data across many devices in the context of space. Additionally, the virtualization of space also enables organizations to integrate other types of information not tied to a device or sensor directly—for instance, the location of people as they move through a store or the location of trash that accumulates in a street. As "Internet of Cameras" in the August 2018 Viewpoints explores, intelligent connected cameras are advancing and becoming more common. Connected cameras examine video, identify key insights, and then report the insights to the cloud rather than passively uploading a video stream. Such insights do not relate specifically to the camera but instead to the space the camera is observing. Digital twins of spaces could provide an ideal way to visualize and process information created by smart cameras.

Scale of Impact

  • Low
  • Medium
  • High
The scale of impact for this topic is: High

Time of Impact

  • Now
  • 5 Years
  • 10 Years
  • 15 Years
The time of impact for this topic is: Now to 10 Years

Opportunities in the following industry areas:

Infrastructure, manufacturing, energy, transportation, health care, software services, sensors

Relevant to the following Explorer Technology Areas: