It’s a tough place to be, the COTS (Commercial Off-The-Shelf) embedded computing space - particularly when serving defense and aerospace customers. Being between the commercial silicon providers and demanding customers presents a number of challenges. It is two moving targets – and it forces us to be disrupters in our own market.
Commercial silicon providers move and innovate at the speed of the commercial enterprise and the consumer. Products are engineered and produced on regular rhythms and become market-obsolete in a relatively short period of time. Defense and aerospace customers want that same latest technology for new program starts - but then require product production support for 10 years or more in many cases. COTS open architecture initiatives have allowed some level of spiral development and backward compatibility, so planned product upgrades can now be part of the overall program development.
So we at GE must consistently design and produce embedded computing product with the latest technology with the objective of securing design wins - and then provide long term support for the product.
There’s a lot of discussion about market, product and technology disrupters. A company needs to be in constant change to stay relevant - and even be our own disrupters in the market. GE does that today across the organization with initiatives to take advantage of GE’s extensive intellectual capital more quickly for the benefit of customers. In the case of defense, aerospace and industrial solution embedded computing we are, to all intents and purposes, driven by commercial silicon developments. This is a large part of our forcing function. We bring this commercial silicon technology to customers in products designed for their intended applications and rugged operating environments - and many times over, this commercial silicon technology disrupts.
It dawned on me at the recent AUVSI (Autonomous Unmanned Vehicle Systems International) conference in Orlando, when discussing our new product development around a GPGPU processer, that I was describing a new technology and product that was disruptive to our existing product. GE is currently developing new high performance embedded computing solutions using NVIDIA’s new Tegra K1 System-on-Chip mobile GPGPU processor. What I was describing was the processor’s step change in performance per watt from previous processor technologies; over 300 GFLOPS in a board design that will consume only 10-15 watts is a rather compelling story. And further, the computing product could also be a fraction of the size (COM Express) of other industry standard boards (3U and 6U Eurocard). Today, we deliver high performance computing solutions in the multi-hundreds of gigaflops to teraflops performance range that are almost ten times that amount of power and five times the volume. That’s disruptive. And as a supplier to this open standards computing market we have no choice. We have to make this computing product and technology available to our customers.
Many of the unmanned systems manufacturers at the Orlando AUVSI event were developing unmanned platforms with size and weight constraints - and they all desired off-the-shelf open standards computing solutions with significant performance capabilities in the smallest and lowest power form factor possible. Their payloads now include multiple camera arrays, multi-modal sensor suites, smaller radars and the requirement to have on-board processing and decision capacity. Size and weight are everything to any aircraft platform, and UAVs in particular.
This new product development fits the bill. But what about product designed just a couple years ago that also potentially applied to these same customers? Have I just moved them from product we spent beau-coup dollars developing over the past couple years to this new capability? Probably so in many cases. Did we have a choice? Probably not.
To our customers they get the best of both worlds. The best and latest technology, and support for the long haul. To us it is constant disruption.