Joint Design Manufacturing Roadmap
Commodity hardware design, build, and manufacturing are at the bottom of the value cycle today. We need a new success framework. Joint…
Commodity hardware design, build, and manufacturing are at the bottom of the value cycle today. We need a new success framework. Joint Business Management meets Joint Design Manufacturing.
Compute and Storage OEM companies tend to follow an ancient cycle.
Design custom hardware to meet specific market requirements
Shift to a commodity strategy that works on “white-box.”
Standardized equipment across platforms to maximize returns
Struggle with new workload and software demands that require specialized hardware designs to keep up. Ever-expanding and more complex use cases overwhelm stale plans. 5G, Edge, Automotive as business models. GPU, FPGA, PCIe Gen 5, and fabrics as technological models.
Specify “helpful” design suggestions to ODM and EMS companies. The intent is to avoid non-recurring engineering, “NRE,” costs while leading vendors on with potential project awards
Give up trying to get vendors on board and create new products. Locking them down as proprietary and complaining about costs
Hyperscale and Cloud companies are now learning the OEM lesson. They turned datacenter models upside-down with commodity deployment strategies and management. Now they require specific hardware designs too. Fighting the same curve as brand ODM’s.
As new technology advances and user requirements drive development, companies find themselves stuck between ODM and traditional OEM models.
Companies who never invested in, or laid off, core research and design development staff struggle to meet new and quickly evolving market demands. Customers pressure ODMs to accept specific design specifications that can’t, or won’t, be attractive to a broad customer base. As small changes turn into unique features, ODM’s rack up higher costs and design errors.
Design errors and long release times come from a protracted and sometimes tricky design specification phase. Customers do extensive market research trying to rule out any Whitebox or generic alternatives. They assemble a team to solve shortcomings and specify changes. Often this ends in new designs based on a Whitebox Frankenstein blueprint they share with ODM partners trying to rush them into production.
ODM’s struggle to force-fit changes on existing platforms or create a new SKUs. Discussions back and forth often create more errors when the ODM tries to force match changes with existing features or functions. Hardware is hard as they say and there are few shortcuts.
Traditional manufactures, CM and EMS, struggle with coin-operated responses that fall short at best or are insulting at worst. Both in response times and knowledge. Unable or unwilling to invest in a tactical engagement team, customers relegate them to arms and legs for manufacturing at low-profit margins. So what’s the answer?
JDM to the rescue. Companies, such as Inspur, who decided to invest in talent, skills, and new business models, showed just how practical and profitable this approach could be. Good interview here: LINK
Inspur took the lead on JDM design models based on existing OEM relationships. Yet, no EMS or CM’s have stepped up tho the challenge. The new Hyperscale-Cloud players need a similar JDM model. But with a twist. Cradle to grave support.
One example of cradle to grave thinking is designing for recycling. It’s the only way to reach a negative carbon footprint. Requirements like this are developing all the time do not fit into old manufacturing models.
Partnership and robust interaction patterns are required. Captured early enough in the design process, companies can reduce complexity. It takes vision and investment. It’s not a burden if staged correctly with the right team leadership.
Hyperscale-Cloud companies understand software models for publishing standards. Hardware does not work the same way. The hardware ecosystem has much more back pressure, given all the parts and pieces required to build a system. From the Foundry to the accessories, everyone has a roadmap and investment model supporting it. Creating ideal hardware to meet customer demands is as much an art as it is a science. CM-EMS companies are likely the best business models to exploit the JDM approach. A hub of sorts that can work broadly across the ecosystem against the backdrop of customer demands.
Hardware takes very concrete steps to prove design, prototype, and manufacturing stability. Customers need to be carried through the process and shown why opening these designs to a broader market is much like creating an open requirements document for software. It has to be interactive and functional across all the technology partners under a given platform.
JDM partners can be a source of knowledge and two-way communication. JDM partnerships require new relationship terms to unlock value. Customers should expect to pay something for this service when available. Vendors should form a tactical execution team to full fill the promise.
The core team, 3–9 people, is sourced from various disciplines. They coordinate a pool within the company across old silos. JDM is a new way of thinking, and companies who grasp it’s power now will prosper just as Inspur has.
Here are critical items for a customer JDM program
Specifications should be a two-way dialog locking in outcomes, not technical details. e.g., power should last after disconnect for 5–15 minutes under full load. The temperature ranges will be between x — y. What’s the max thermal envelope for the platform.
A PRD should come from the specifications phase, not the other way around.
Questions and clarifications should not be seen as a chargeable event by the manufacturer or a freebie by the customer. A selling point for the partnership agreements
Manufactures should drive the value of transparency and reuse within the customer base to find solutions. Define precisely what aspects of the design are critical to the customer while creating a useful market platform.
Customers should invest in training. Side by the side JDM sessions should be created to educate partners on cultural values. The partner’s core team should hold customer interaction profiles to train silos throughout the entire customer engagement.
Executive champions should be stable on both sides of the relationship. Precise and compelling lines of communication designed to be a positive escalation path vs. the current blame game models. Solution-based vs. the regular, top-down punishment model it is today
Partnership with a “win some lose some” value model
While hardware is hard, the software is complicated. Especially lower down in the stack. Low-level software and the ability to manage it should be in every excellent JDM partnership opportunity. A separate contract for these services is required.
Cradle to grave JDM partnerships doesn't fit into standard contract agreements today. New relationships and business models must be created. No one can go it alone in this market. It’s time to develop joint business management to support joint design manufacturing.