Lifecycle management approach

Commercial technology and components are often changing much more frequently than the lifecycle expected of extended lifecycle OEM compute engines. Computer components such as chipsets, processors, video processors and optical drives often have product lifecycle availability of 6-months or less.

To leverage commercial technology in an OEM compute engine with a longer lifecycle than the components within it are typically available, the OEM compute engine manufacturer must use one of several business models to fill this availability gap:

Model 1: The OEM purchases enough identical systems and stores them, using them through the life of the program.

This approach is easy for the computer manufacturer (big lump order) but is an economic challenge for the customer, as they have to buy, pay for and store several years worth of product. The computer manufacturer can also manufacture larger batches of products. The customer also takes-on the risk of the accuracy of their initial forecast (what happens if actual demand exceeds the units purchased?)

Model 2: The OEM compute engine manufacturer or their distributor manufactures enough identical systems and stores them, using them through the life of the program. Generally this is facilitated by a blanket PO or order commitment from the OEM customer.

This approach places some burden on the computer manufacturer or distributor (the cost of the systems in inventory). It also places a risk factor on the OEM customer by requiring that they commit to volumes at the very beginning of the program, when their forecast accuracy may not be precise. Commitments that are too large or too small will carry an economic burden on the OEM customer.

Model 3: The OEM compute engine manufacturer buys enough components at the program start to enable manufacturing of identical systems through the length of the project.

This approach is harder for the computer manufacturer, as they must source, carry and manage a significant inventory of parts, and perhaps run smaller custom manufacturing batches of systems. The customer has less economic burden (since they buy systems as they need them) but still carries the risk of the accuracy of their initial forecast.

Model 4: The OEM compute engine manufacturer has a product longevity “system” that manages the process throughout the lifecycle of the product, and minimizes the impact on the OEM customer throughout the project.

This approach places the most significant burden on the computer manufacturer, as they must select technologies, components and vendors carefully, and they must be continuously monitoring the availability of components during the product lifecycle. The manufacturer must also carry an inventory of parts as parts go EOL. This approach has the least impact on the OEM customer as they do not carry the economic cost of carrying significant inventory, and their forecast accuracy can be significantly increased by making those forecasts well into the program.

Dedicated Computing uses a layered product longevity “system” to bridge the gap between the platform longevity needs of customers, and the variability of the commercial component and subsystem marketplace.

The first and most fundamental layer in the methodology is to architect systems that utilize technologies that will be accepted by the market, will be in volume production for many years, and will not go obsolete quickly due to lack of market demand. The ability to select these technologies is based on a significant effort to keep current with technologies, trends and roadmaps. This efforts often revolves around whether to jump on a new technology early, or wait until market acceptance has been determined.

The second layer in the methodology is to select components that have been identified by their manufacturers’ as being long-life components. Many component manufacturers (such as processors, chipsets, etc.) select a subset of their offering to make available for a much longer period of time. Selecting these components ensures that subsystem (board) manufacturers will be able to source the components they need to manufacture the subsystem.

The third layer in the methodology is to select subsystems (such as boards, drives, etc.) that have been identified by their manufacturer as having a long product availability. The longer the subsystems are being manufactured reduces the amount of EOL inventory that may need to be held near the end of the program.

The final layer of the methodology is for Dedicated Computing to source and stock any of the particular components or subsystems that do go EOL prior to the end of the stated product lifecycle. EOL quantities are calculated based on input from customers, and since this generally occurs later in a program lifecycle, more accurate data is generally available to the customer.

Finally, outside the scope of the availability of a particular OEM compute engine is the transition to a next-generation system. Dedicated Computing manages the process of identifying next-generation hardware that creates the maximum possible overlap between the mature system and the next-generation system. This enables the OEM to transition to the next-generation system on their marketing schedule, minimize EOL inventory and risk-reduce the transition process.