Standard processors are highly susceptible to attack. We've lived with that fact for decades, stuck with what was instead of moving towards what could be. Why has no major processor vendor begun to develop an inherently secure processor that can withstand the exploitation of common software vulnerabilities? Probably because any new processor that resists the root causes of modern vulnerabilities would require a brand new operating system, new programming languages, and applications to be rewritten from scratch.
Except, that reasoning is no longer true. Key DARPA innovations derived from the CRASH program that address the root cause of computer insecurity at the processor level can be applied to commodity processors. In the world that we envision, modern vulnerabilities are virtually eliminated beneath the OS by differentiating data from instruction and enforcing memory use constraints within the instruction pipeline. This talk introduces our first step to inherent security, where, by using the open source RISC-V processor as our foundation, we intend to provide provable resilience to common vulnerabilities by July 1, 2016. We'll chat about our ambitious plan for the next year, how we expect to support standard operating systems, and how the public can participate.