We had a chance this week to hear how IBM Research goes about its mission of inventing the future. Instead of hiring the brightest scientists and locking them in an ivory tower with like-minded people, the organization strives to embrace collaboration and a full-stack scope of solutions. The different perspectives and contributions from clients, partners, and other research organizations can combine to produce the hardware, software, and systems needed to solve a problem with a superior solution.
The Benefits of the Full-Stack Approach
If you are a chip designer, the solution to a problem is probably a new chip. But if you are a software designer, you sit down and start writing code that will run on an existing chip. But if you are IBM Research scientist, you get people together from all relevant domains and collaborate on a full solution to a problem, co-designing across various disciplines. By co-designing with clients, partners, and technology pillars, a comprehensive approach will produce superior results compared to a piece-meal solution.
A good example is how IBM Research has addressed the needs for in-line or real-time AI to prevent credit card fraud, ensure security, and validate transactions. IBM combined semiconductor design for reduced precision math operations with software, firmware, SoC design, and full systems engineering to solve these and other customer needs in business-critical transaction processing.
Some people still think IBM got out of the hardware business. Yes, they exited hardware businesses that were commoditizing, like x86-based PCs and servers. But they didn’t exit the high-value (and margin) business that remain above the commodity fray in the market place, and they remain in these business precisely to enable a full-stack solution to hard-to-solve problems.
In this video, IBM VP of Hybrid Cloud, IBM Research, Mukesh Khare explains more.
IBM Research stands at the crossroads of technologies that can solve difficult problems, playing a pivotal role in fundamental technologies such as semiconductors, materials, manufacturing, quantum, hybrid cloud and analog computing. By adopting co-design principles, projects across IBM Research address the full stack of technology needed to meet challenging problems holistically.