ARM isn’t a semiconductor company in the traditional sense. It is an intellectual property provider with a significant presence in processing IP across a wide range of applications from mobile phones, to IoT devices, to the core of infrastructure.
“We license our IP and variants of that IP to semiconductor companies who embed that into chips to put their own value around that,” said Mohamed Awad, VP of marketing for ARM’s infrastructure business.
Instead of manufacturing chips, ARM provides its instruction set architecture to its customers — such as Qualcomm, Broadcom, NXP, and Marvell — and those companies manufacture the chips.
While some chips, such as Intel’s x86, use complex instruction set computing (CISC) architecture, ARM chips use reduced instruction set computing (RISC).
“At the lowest level there are a set of instructions, which the processor executes,” said Awad. “The speed at which the processor might execute those instructions, the number of instructions, how the processor is configured, what features it might have — those will all differ. But the baseline instruction set is RISC, low complexity instructions.”
One of the big benefits of RISC architectures is their low power consumption. Because of their reduced power needs, ARM-based chips have achieved wide adoption in mobile and IoT devices.
“Our bread and butter for years was mobile phone chips,” said Bob Monkman, director of software strategy for ARM’s network infrastructure group, in an earlier interview with SDxCentral. “That’s where we made a name for ourselves.”
The company makes three broad categories of chips: “A,” “R,” and “M.”
“A is an application processor that tends to be among higher performance things,” said Phil Bourekas, director of marketing for networking infrastructure at ARM. “For the most part, the class of stuff in a server in enterprise networking equipment doing control plane processing — those are going to be A series processors.” R chips are used for real-time processing. And M chips are used in micro-controllers for such things as IoT endpoints.
Awad reiterated that it’s important to remember that Arm doesn’t actually build the chips. “We build the IP that our partners configure. Even though the underlying core is the same, the design variances are important.”
Although chips for mobile devices have been ARM’s most successful business, the company has also expanded into servers and networking infrastructure.
Of this business, Awad said, “We have an active pipeline of customers and partners who are building products based on ARM. Among the names are Marvell’s Armada chips and Cavium’s ThunderX2 chips. (Marvell recently acquired Cavium for $6 billion). Awad also named NXP’s Layerscape chips, and Qualcomm’s Centriq line.
In August, ARM announced it was establishing an internet of things (IoT) platform named Pelion. The platform will include IoT connectivity technology, device management, and data management. ARM is cobbling the platform together with technologies it acquired from its purchase of Stream Technologies along with its ARM Mbed Cloud. And it will also integrate the data management technologies from Treasure Data, which ARM recently acquired.
SDxCentral asked how Pelion relates to ARM’s core competencies as a chip designer. In an email, an Arm spokesperson said, “The Pelion IoT Platform definitely complements our chip design business.” He cited ARM’s IoT device IP in which ARM helps designers to build their system on a chip (SoC) for IoT devices.
In addition, ARM’s Kigen provides a portfolio of Subscriber Identity Module (SIM) solutions for secure identity of IoT devices.
And the company’s Platform Security Architecture is a contribution from ARM to the IoT ecosystem to build more secure IoT devices.
Finally, ARM’s Mbed OS is a free, open-source embedded operating system that includes all the necessary features to facilitate the development of IoT connected products.
ARM is working on a set of 5G initiatives, especially around the processing of increased amounts of data.
“5G will have devices from IoT endpoints all the way to mobile devices, which will generate more data that will need to be processed by the mobile edge, and in the core,” said Awad. “We are absolutely working on a set of initiatives around that. And I don’t think we can underestimate the impact that’s going to have on the data center and the network infrastructure.”
He added, “Obviously 5G is a huge deal, not just for networking infrastructure, but communications at large. It creates this massive new pipe characterized by lower latency and greater bandwidth. In many ways we are leading in some of these areas. We have to develop the IP. That takes time. Then our partners have to take that IP. We’re thinking about this two or three years ahead.”
Photo Copyright: almoond / 123RF Stock Photo