Arm is expanding its Total Solutions for IoT full-stack product development suite, which it announced last October, and augmenting its IoT processor offerings with upgraded performance for IoT-related machine learning and security capabilities.
The need for increased focus on these two particular areas can hardly be overstated, as the integration of IoT and AI-related technologies like machine learning is leading to the emergence of an “AIoT” solutions segment focused on processing all the machine learning data coming through IoT sensor networks. At the same time IoT security attack numbers have exploded, with a recent report from security vendor SAM Seamless reporting that there were as many as 1 billion cybersecurity attacks on IoT devices last year.
The new Arm Cortex-M85 processor was described by Mohamed Awad, vice president of IoT and Embedded at Arm, during a media briefing, as an “ideal choice for applications that have additional performance and security requirements.” He added, “It starts off by adding enhanced micro architecture features for increased scalar performance, and then it adds the power from Arm Helium [vector extension] technology to enable support for machine learning use cases as demanding as high-fidelity audio processing without the need for a DSP.”
The M-85’s scalar performance is 30% higher than Arm’s Cortex-M7 processor, Awad said.
Awad also said that as malicious attacks on IoT systems become more frequent and varied, the M-85 brings support for Arm’s TrustZone enforced hardware isolation technology, as well as being the first Cortex processor to integrate the Point Authentication and Branch Target Identification, the latter being a new architectural feature with enhanced software attack threat mitigation that “simplifies and accelerates the developer journey to PSA Certified Level 2,” a security baseline for IoT deployments, Awad said.
Among other IoT announcements Arm made this week, Awad said the company is expanding its Arm Virtual Hardware library, which is intended to help developers and software companies more easily and affordably develop their own IoT solutions for smart home devices and other IoT by leveraging reference designs based on Arm hardware.
The Arm Virtual Hardware offering is now available for:
-
Three Corstone platforms: Corstone-300, Corstone-310, and Corstone-1000. Support for Corstone-1000 extends the scope of Arm Virtual Hardware to applications requiring rich operating system support. Arm Virtual Hardware Corstone is available as Amazon Machine Image (AMI) on AWS Marketplace.
-
Seven current Cortex-M processors: Cortex-M0, Cortex-M0+, Cortex-M3, Cortex-M4, Cortex-M7, Cortex-M23 and Cortex-M33. Arm Virtual Hardware processors are available as Amazon Machine Image (AMI) on AWS Marketplace and locally through Keil MDK-Professional.
-
Three complete IoT development boards in partnership with Raspberry Pi, NXP and ST, allowing for the seamless transfer of software from virtual to physical hardware. Arm Virtual Hardware third party hardware is available today as a private beta for the Raspberry Pi Model 4, NXP i.MX 8M Arm Cortex Complex and the STMicroelectronics Arm Model for the STM32U5 Discovery board.