As we develop electronics in early 2014, the battle between processor architectures is raging in all spaces, from deeply embedded through mobile to servers. Choosing the right ecosystem partners is crucially important, and today's announcement of an extended partnership between ARM and Cadence marks important steps towards increasing productivity for ARM-based designs.
About two years back, I had tried to make sense of what I perceived as chaos at the Embedded Systems Conference, pointing to processor architectures like ARM's and Intel's as the keystones for the ecosystems to which we contribute development tools and peripheral IP. I have just returned from the Embedded World conference in Nuremberg - several times bigger than ESC in the US - and the picture has not become any less fragmented. ARM celebrated the shipment of 50 billion devices at their booth on Wednesday of Embedded World - they are certainly a key ecosystem partner for Cadence. In a brief video interview two years ago, I actually talked about how we work together with ARM in the systems domain.
And dependencies have become even closer since then. Consider the illustration associated with this post, outlining a chain of electronics found in the Internet of Things (IoT). Devices create data, like in my case the sleep and step tracker on my wrist. Data gets collected in some type of hub - my phone or my computer. Other hubs are my car and potentially my living room devices (for home-related items like my solar panels and alarm system). The hub collects data and - when available - uploads data to the cloud. And in the cloud my data becomes incredibly valuable as a source of "big data" analytics. I am not quite sure what would be worse - my Mom knowing the low number of steps I may be making on a lazy day, or my health insurance getting their hands on that data to know my work-out habits. My conscience is pretty clean on both, but nonetheless ...
This chain shows how the value is created not by the individual components but, instead, by the combination of them, from device though hub to the cloud that holds the data that is analyzed! The server space that holds the data to be analyzed is the traditional x86 space into which ARM has entered with the ARMv8 64-bit architecture. The mobile space - the hub - is the space that ARM dominates and into which Intel has been trying to enter for quite some time with the Atom architecture. Finally, the traditional deeply embedded space of devices and sensors is the space into which ARM has entered with the Cortex®-M and Cortex-R architectures.
The announcement we made today at CDNLive talks to our partnership with ARM in three areas, focused around the v8 architecture.
As it relates to the devices measuring the data in the first place, we have already talked about how ARM and Cadence work together to enable Cortex-M based systems. We also showed at Embedded World last year how software on a Cortex-M processor can be simulated together with the analog/mixed-signal components.
Enabling developments for the Internet of Things requires strong ecosystems, and with today's announcement, the ARM ecosystem just became stronger!