Nonetheless, all such devices share the same bottom line and all must vie for bigger profit margins and meet tighter deadlines while squeezing more specialized features into every new release to maintain or increase their market share. Thanks to the global and fiercely competitive markets of today, products across the electronic gamut are evolving at a breakneck pace. Consumers are more sophisticated and more connected, demanding much more from their mobile phones, their networks, their cars and all the gizmos in their lives. In turn, these demands put a severe load on the product and its production alike. The product load exerts pressure on the hardware and software platform in terms of performance, security, flexibility, and interoperability. These devices are being asked to deliver a myriad of multimedia and Internet protocols faster than ever without compromising the trustworthiness of the device or the identity of its user. The production load on the other hand shows up in the development life cycle that mandates reuse wherever possible and faster turnaround times to achieve the ultimate goals of shrinking the bill of material (BOM) and time-to-market (TTM).

Advancement in multicore chip architectures represents a revolutionary response to the product pressures at the hardware level. Such architectures promise to allow system designers to produce next-generation systems that easily scale to satisfy the onslaught of feature and computational demands, packing more into a smaller consolidated footprint at a cheaper price. This would then eradicate a good portion of the pressures except that a whole slew of complications arise in the software plane with the introduction of multicore. The software that has been written to run on billions of embedded and connected devices in use today does not fit the parallel processing paradigm that multicore brings about. The majority of the legacy software in existence today is neither multicore-optimized nor is it multicore-safe.

Enter the next wave of trusted real-time virtualization. Virtualization in this scenario plays a unique role in alleviating these pains by decoupling the hardware and software planes in a flexible way that presents the operating systems and software applications with an optimum configuration of the hardware underneath. This virtualization enables competing operating system environments to coexist on a single hardware chip while segmenting or virtualizing the hardware capabilities in a manner that is usable by the software without imposing any immediate changes on the latter. This can be a wild mix of operating systems, including real-time (RTOS) ones vying for deterministic access to system events next to general purpose (GPOS) ones that are able to tolerate jitter and minor delays. This also enables designers to slowly absorb the impact of the new multicore hardware reality on their code by way of containment, where legacy code is contained in a virtual machine (VM) that is presented with a single core. Newer code that is capable of harnessing the power of multicore can run in a distinct VM spread across the available cores or even share already-used ones for that matter. Legacy code is hence run unaltered and readily side-by-side next to newer open and rich environments. This provides system designers with unparalleled power to innovate while reusing existing intellectual property at no added costly engineering revalidation cycles.

Another use case for this type of virtualization has been to enforce the quality of service (QoS) and provide secure isolation of mission-critical services from other less-critical functionality on the same device. In the instance of handsets, it is very beneficial to secure the software that places voice calls on the wireless network against failures that might be caused by unwanted tampering in the GPOS on the device, intentional or otherwise. This is best done by having two VMs run side by side with the voice-handling software isolated by the virtual machine manager (VMM) from the rest of the device software. A reboot of the GPOS will not incur a service disruption or even affect the quality of the voice conversation, which would continue uninterrupted while the system comes back online.

The idea of having device software compartmentalized is very enticing indeed. Interoperability and flexibility, another two important product pressures, can be extended using real-time virtualization. As an example concerning connected devices from handhelds to static access points such as WiMax and Femto cells, various networking stacks can now be hosted in separate VMs providing devices with a number of options to connect with an outside network.

From the operational pressures perspective, the specialization and efficiency of this next-generation virtualization unlocks the power of newly available hardware and open source software without incurring the traditional costs of porting existing platforms. The risk of adopting newer technologies is then mitigated and can be phased out over time, extending the earning power and lifetime of existing products while adding differentiation. Factor into this equation the classical benefits of being able to create open environments, allow for mobility of applications, and reduce the cost and project timeline and this technology becomes much more attractive across a wider range of device software, beyond the reach of the server and desktop domain.