For large-scale network deployments, approximately 30% of ASP is tied to the software stack, which also accounts for 80% of an OEM's R&D. When it's time for NGN, it doesn't make sense to throw away a successful product. Oftentimes companies have to recover their R&D investment through multiple generations of products. In the meantime, hardware continues to evolve. To remain competitive, OEMs must migrate existing product lines to the latest silicon. Migrating applications to new hardware with minimal modification can extend the life of those applications while increasing their performance.
An OEM's success in the device market is directly tied to its ability to break the inertia of its application software and rapidly port it to newer, faster, cheaper hardware. MontaVista has multiple solutions for creating a virtualized environment within the hosting Linux operating system. One of the primary use cases for these virtualized environments is to facilitate migration of applications from legacy hardware to new hardware. In addition to performance, benefits can include faster time to market for updates and the reduced cost of a Linux environment.
The migration effort itself can be significantly improved through the virtualization capabilities in MontaVista's platform products. Options include full-system or lightweight virtualization:
In full-system virtualization, using CGE or CloudStructure as a network abstraction platform, the entire legacy OS environment can run as a virtual appliance, allowing a staged porting approach wherein the legacy environment is preserved while new features are added to work with the advanced APIs on the hosting Linux environment. Most new architectures have legacy and new applications all running as virtual appliances to future-proof next-gen products and speed time to market.
In lightweight virtualization, a container can be created in the CGE environment, where parts of the legacy application can run with guaranteed CPU time and memory availability and minimal latency. This setup can be further enhanced by making use of CGE features that allow near bare-metal performance for dedicated cores using multi-core hardware.