Managing the military component obsolescence paradox: When new performance levels are needed after EOL
Long-term delivery assurance, replace or upgrade programs, or shifting architectures are each an effective means of adding longevity to established legacy systems, and primarily depend on evolving performance demands. Competitive design options allow proactivity in managing obsolescence of critical applications - while focusing on improving performance.
Obsolescence management can be a tough challenge for military systems designers, but is always an essential consideration in plotting a successful product life-cycle policy. When critical components are no longer available and new performance features are continually required, other assurances are necessary in order to preserve program integrity. Adequate preparation for effective product life-cycle management, for example, ensures predictable product consistency and longevity. At a minimum, it gives mil/aero customers the ability to plan for the normal lifetime of a product.
Yet despite a manufacturer’s best efforts, there might be a need to modify a product. This might be the result of a component prematurely becoming obsolete, an enhancement that is desired by the market, or a correction to a problem. Because there might be the need to modify a product or end its lifetime prematurely, product life-cycle management can help designers respond efficiently and effectively to unexpected changes in a product’s life.
Early notification, planning, and options are all imperative to determining the best design path to mitigate obsolescence. Updating components might migrate the system out of obsolescence; however, it might be more imperative to address requirements for integrating new technologies at the same time. Issues such as Size, Weight, and Power (SWaP), system performance improvements, connectivity, scalability, reliability, and enhanced ruggedness might need to drive the process.
Essentials of early notification
Product Marketing Information (PMI) ideally provides one-year notification (and a minimum six-month notice) that a component is being designated as EOL. Designers may elect for a last-time buy to support the product through its program life cycle, in which case the manufacturer can take steps to secure product as required. Included in the PMI are options, including recommendations for additional boards or components that will address the performance requirement.
Designers may choose to stay with an upgrade path supported for initial long life and then further extended through the enhanced next-generation components recommended via the PMI. However, when these upgrade paths are no longer appropriate given new performance demands on the application in question, designers may choose to work with the manufacturer for an alternative solution, which might be best handled by a platform change.
Options and planning make all the difference
Planning product lifetimes starts as early as the development phase, dependent upon choices that affect the basic overall product life cycle (Figure 1). Components are selected based on features, market acceptance, and the approval of the vendor with the intent of ensuring long-term supply. As products mature, they are manufactured only for existing customers or for new designs that do not last longer than the product life cycle. The mature phase is usually two to three years, but will continue as long as there is sufficient demand or until component obsolescence makes it impossible to build the product.
Typically, a product of equal performance and features is available as a replacement for the product at EOL status. These parallel replacement product versions are developed as a part of a normal product redesign cycle to take advantage of newer or more economical technology. When a particular product family member enters EOL, other higher-performance products are available (or planned) within the product family and can be used as an upgrade replacement. Alternatively, many product families have members that are quite similar in performance and features, but differ in form factor. If the mechanical integration issues can be easily overcome, alternative replacements might even be a desirable path to capture new technology or other benefits, such as a reduction in size.
In these scenarios, SWaP might need to be decreased in a particular integrated system in order to decrease SWaP levels within the overall system. This approach is common when OEMs are working to manage obsolescence of existing platforms while introducing other systems elsewhere in the platform. Shipboard and ground vehicle applications are common areas for these design issues, with designers working to pack more functionality into a finite space that can only be extended by reducing the footprint of existing systems. As combat environments evolve, SWaP reduction can improve troop safety simply by enabling a more streamlined deployment with long-deployed systems.
Managing obsolescence with architecture upgrades
A deeper level of upgrade might be considered, for example, if there is a need to increase performance based on new software capabilities established since the system’s initial design or deployment. Additional features might now be required, such as higher CPU performance or increased memory. In this scenario, physical requirements might be flexible enough to allow a form factor change as warranted. For instance, an existing VME system that needs to incorporate higher-bandwidth technologies might need to evolve to VPX, but that requires changes to be made in the backplane and all system cards; this is significantly different from a simple CPU card upgrade and often requires greater design expertise.
The VPX architecture represents a dramatic shift from VME communication protocols, with signals moving across Serial RapidIO, Gigabit Ethernet, or PCI Express instead of the PCI or VMEbus. New High-Performance Embedded Computing (HPEC) platforms that are VPX-based supercomputer-like systems are gaining ground as a suitable option – providing massive processing power for compute-intensive DSP-based systems and allowing high-speed socket-based communication between blades by using multiple switched fabric interconnects within the backplane.
VPX replaces the bus with a network-based protocol, which typically demands application software retooling, and in turn drives military designers to consider 3U CompactPCI as a viable upgrade alternative for 6U VME-based systems. Its reduced form factor meets SWaP goals, and CompactPCI provides a well-established parallel bus standard, which provides a cost-effective modular computing platform that more closely resembles VME in terms of how application software recognizes the hardware.
Migrating to x86 can be a viable performance option
The cost of an architectural redesign might be too high, or specialized I/O boards might be difficult to replace. In these cases, designers can manage obsolescence by migrating to a new processor architecture that enables lower power and higher performance. For example, a VME design could transition from a PowerPC architecture to x86 by integrating current components that support the latest Intel processors. In this example, designing systems around 6U VME boards allows the final system to span different CPU architectures, which helps reduce risk and development time. This is essential as upgraded designs typically need to be put in place quickly with minimum risk to the overall system or application.
Further, designers can shift easily from 6U VME implementation to x86 in the 3U CompactPCI and 3U VPX platforms. Many designers are taking the opportunity to upgrade the backplane in the migration process, gaining bandwidth and performance features by moving to CompactPCI and VPX. In making this transition, users can go from 2.5 Gbps data transfer for VME to 4.2 Gbps for CompactPCI or up to 10 Gbps for VPX. This represents a change in design thinking; although they cannot put as much hardware on each card (6U to 3U form factor change), designers have more space at the overall system level resulting from moving to a smaller form factor. Processors with faster clock rates and increased power-to-performance ratios mean they can pack the same performance in a much smaller volume.
Overall though, Intel x86 architecture is form factor-agnostic and applies readily to any number of the military’s favorite platforms. The end result for military designers is that they can boost performance in a standards-based, multicore platform that is able to meet highly demanding signal- and data-processing requirements, an ideal fit for applications onboard submarines, naval ships, aircraft systems, and ground vehicles. For example, today’s products based on third generation Intel Core processors enhance available small form factor solutions and provide up to 20 percent greater computing power and up to 40 percent increased performance per watt compared to designs based on previous generations.
Today’s new x86-based embedded computing platforms combined with FPGAs enable another new realm of applications – providing highly adaptablefeature options for designs that have previously been restricted due to lack of interface or I/O support (Figure 2). By understanding the collective advantages of this approach, designers can reduce Bill of Materials (BOM) costs and maintain long-term availability with legacy interfaces and dedicated hardware-based I/O. Most importantly for legacy systems facing upgrade or redesign, there is now a bridge to tap into the latest processor enhancements such as graphics media acceleration, hyperthreading, and virtualization for greater success in matching exacting requirements. The FPGA solution allows the designer the ability to replace legacy or obsolete I/O and still maintain system integrity. This might require some initial software development if an existing IP core solution is not available. This is a significant advancement in bridging newer technologies with older systems implemented in the military market.
Budget, performance most vital in determining design path
Defense budgets are tight and the pressure is on. System deployments are being extended years longer than originally anticipated even while performance expectations are higher than ever. Designers of today’s systems are challenged with getting creative – understanding evolving standardized platforms and finding the best embedded computing options to keep military applications and systems performing up to and beyond battlefield expectations.
Performance and reliability are essential: Older systems must be migrated and consistently enhanced to meet increasing levels of sophisticated data sharing, ruggedness, and performance. Designers approach obsolescence from several perspectives, each of which impacts their platform choice. Each approach might not be exclusively appropriate for a certain application or deployed environment, and designers will find it necessary to make trade-offs between performance, development time, cost, and legacy compatibility.
Based on costs and DoD budget requirements, many large, legacy military programs consider remaining in VME the most viable option – replacing legacy VME chassis, I/O cards, and software with products that now offer improved availability, performance, and features based on x86 architectures. In turn, many embedded computing suppliers are competing with this mandate, developing high-performance VPX and CompactPCI systems in parallel that deliver a range of compatible options designed for pure performance and reliability. Most importantly, system designers have a growing slate of competitive design options that allow them to be proactive in managing obsolescence of critical applications – while improving performance.
Kontron America 858-677-0877 www.kontron.com