Arm Community
Arm Community
  • Site
  • User
  • Site
  • Search
  • User
  • Groups
    • Arm Research
    • DesignStart
    • Education Hub
    • Graphics and Gaming
    • High Performance Computing
    • Innovation
    • Multimedia
    • Open Source Software and Platforms
    • Physical
    • Processors
    • Security
    • System
    • Software Tools
    • TrustZone for Armv8-M
    • 中文社区
  • Blog
    • Announcements
    • Artificial Intelligence
    • Automotive
    • Healthcare
    • HPC
    • Infrastructure
    • Innovation
    • Internet of Things
    • Machine Learning
    • Mobile
    • Smart Homes
    • Wearables
  • Forums
    • All developer forums
    • IP Product forums
    • Tool & Software forums
  • Support
    • Open a support case
    • Documentation
    • Downloads
    • Training
    • Arm Approved program
    • Arm Design Reviews
  • Community Help
  • More
  • Cancel
DesignStart
  • Developer Community
  • IP Products
  • Processors
  • DesignStart
  • Jump...
  • Cancel
DesignStart
DesignStart blog Why is SoC Design so hard?
  • Blog
  • Forum
  • Videos & Files
  • Members
  • Mentions
  • Sub-Groups
  • Tags
  • Jump...
  • Cancel
  • New
DesignStart requires membership for participation - click to join
More blogs in DesignStart
  • DesignStart blog

Tags
  • Software
  • Cortex-M0
  • Cortex-M1
  • partner
  • Cortex-A5
  • Design & services
  • IP Integration
  • Cortex-M3
  • IP Blocks
  • DesignStart
  • Embedded Systems Solutions
  • Subsystem for Embedded
  • Cortex-M4
  • Software Development
Actions
  • RSS
  • More
  • Cancel
Related blog posts
Related forum threads

Why is SoC Design so hard?

Jim
Jim
March 1, 2017

The exploding IoT (Internet of Things) chip market is expected to reach 20 billion devices by 2020 by many predictions. The majority of these will NOT be wearables. 

 

The IoT market will be driven by Data and Subscription business models. THOUSANDS of these companies popping up will need THOUSANDS of unique, specialized devices. Who's going to design all these chips? Not the traditional IC designers. It's going to be software engineers, systems designers, FPGA designers, college students… you name it. 

 

If you are not a traditional IC or ASIC designer, then you will likely be overwhelmed and have major anxiety when tasked with designing and producing an IoT chip.  

 

This brings up several questions:

  • How are these "new" IoT designers going to find information? Where do they look?
  • How do you train these “new” IoT designers and comfort their anxiety?
  • How do you calm the "sticker shock" IoT company’s managers will have over the exorbitant cost of tools to do the job?
  • How will they find the expertise they need?

 

I believe the solution will be that foundries, design houses and IP companies will have to join forces to produce easy to use reference designs using a higher level approach similar to Raspberry Pi and Arduino.   

 

This means:

  • Modular Architectures specific to IoT Devices
  • Class Libraries for hardware such as Analog Models and Digital Subsystems
  • Software Abstractions using standard API’s and HAL’s (Hardware Adaption Layers) such as CMSIS.

 

Fortunately most IoT device architectures are very similar to the typical Cortex Mx or Cortex Ax architecture shown below.

 

 

Maybe the problem is not that hard after all. The industry has to re-think how they service this new set of IoT designers coming very soon. The ARM Approved Design Partners can help.

 

In my opinion, ARM is the defacto vendor supplying the CPUs for most embedded chips and IoT is certainly no exception.  ARM provides the horsepower with their Cortex-M series and A series of CPUs.  The best part for IoT devices is that they have “TrustZone” for security. “TrustZone” efficiently separates the most secure parts of the chip from the user portions to effectively create a silicon “firewall” for the best security.

 

In addition to security, ARM has created very good “modular” software support with their  CMSIS API and Hardware abstraction layers.  This supports the higher level of abstraction that is needed to enable the thousands of non-chip or non-ASIC engineers soon to hit the design community. In other words, they have to know little about the actual “on the metal” hardware if the IP vendors will provide the CMSIS support.

 

You can see the rest of my thoughts from my presentation at the REUSE2016 conference by going to

https://socsolutions.com/Presentations/REUSE2016_post.pdf

 

Thanks for reading.  I would like to hear your thoughts.

Jim

Anonymous
DesignStart blog
  • Pre-silicon Software Development with Arm Models for Cortex-A5 DesignStart

    Rob Kaye
    Rob Kaye
    Cortex-A5 DesignStart: Take advantage of the fast simulation models of the CPU with a pre-validated, configurable reference (subsystem) design to reduce device complexity and speed-up software development…
    • September 23, 2019
  • Case study: RelChip turns up the heat with Arm DesignStart

    Chris Shore
    Chris Shore
    Learn how RelChip developed an Arm Cortex-M based microcontroller with the widest operating temperature range in the world to be used for applications in extreme conditions, from geothermal to aerospace…
    • May 20, 2019
  • Case study: accelerated chip design for drones and cameras with Arm DesignStart

    Chris Shore
    Chris Shore
    Find out how NeoWine turned to the Arm DesignStart program to create a secure encryption ASIC and high performance MCU for drones, based on the Arm Cortex-M0 and Cortex-M3 processors.
    • May 13, 2019