Connected products is a passion, which began by connecting early personal computers to legacy lab equipment in the late 70s. The demand for connecting the equipment came from local university labs and allowed the creation of a wide range of hardware and software solutions to unlock data from the legacy equipment.
Decades later we are doing the same thing, connecting to legacy products to collect and derive value from the data.
Later I had an amazing opportunity at Apple Computer in the Advance Technology Group. Back then we didn’t call it IoT, we adopted “ubiquitous computing” as the term. Among many fun projects, one of my favorites was developing new hand-held and embedded computer product concepts, networking them together with very early wireless technology. The work influenced several Apple spin-outs, General Magic, Kaleida, and Echelon lonworks.
After Apple, my love of connected products did not end, I founded a Palo Alto Incubator to develop connected product companies. One of my favorite developed the concept of “microserver” and “picoserver”. The picoserver concept didn’t fly as there was not enough ethernet deployment to support networked edge device for measuring and controlling the real world. The microserver however delivered on the concept of one web service per hardware server product, making it much easier to build out the next scalable dot com site. The microserver company was Cobalt and it sold to Sun for $2B.
Ask someone what IoT is and they will describe a system from their point of view, built from their experience. Connected product value Is in the complete system - from where the data is captured at the edge all the way to the human experience to derive some value.
|
|
Tom Gilley |
Tom Gilley has been connecting things to computers since his first job in the 1970s, and since his early development work at Apple, he’s been a serial entrepreneur, creating and incubating a broad range of companies including Wot.io, FutureTel, PicoStar and VidaVee. It’s not an exaggeration to say he’s something of a Renaissance Man when it comes to connected industry. Our conversation highlighted the IoT data value locked up in the industrial companies and the challenges startups face in building connected products.
What I bring to IOT is a blend of the technical and business experience. I like to listen and be a sounding board, engaging with the individuals of an organization to understand the system. Being able to listen and engage with the engineers as well as the executives is what I enjoy. That’s my passion, and that’s what you need to do in both startups and large corporations for success.
I have a firm belief that there is a lot of IoT data value locked up in the industrial companies. Industrials are starting to reinvent themselves as data companies, leveraging the products that they already have deployed and on the line card. Xylem is a fitting example, they have hundreds of monitoring and control products in their line card, many products categorized as Remote Terminal Units (RTU), Programmable Logic Controllers (PLC) and Human Machine Interfaces (HMI) communicate with industrial systems over MODBUS, CAN, PROFIBUS, LONWORKS to name just a few. Many of these products are installed without connectivity to back-end systems to derive value from the device data.
Xylem recently acquired Sensus for $1.7bn in cash. One of the strategic rationales for this transaction is acceleration of connected products through the adoption of Sensus FlexNet connectivity solutions for Xylem products. The adoption of Sensus FlexNet for end to end connectivity will unlock the value of the Xylem product line and generate new revenues from the newly connected product data.
Another area of opportunity growth is in enterprise software solution providers integrating and creating business value from connected product data. For example, a large hotel chain has growing list of connected products; refrigerators, locks, HVAC, security, and paper towel dispensers, all for operational awareness. The connected product data is locked up in isolated operational platforms that can be orchestrated and made available to the existing enterprise software for business decisions. Large enterprises want to leverage the investment they’ve already made in enterprise software for business decisions, connected product software companies should recognize this and adapt to the existing enterprise systems.
A common issue we see are the founders have a great idea but don’t have industrial domain expertise to build the trust in the market. Connected products that target industrial enterprises will need to build trust and demonstrate solid execution. A good technique is to partner with advisors with domain experience to assist in building the market trust.
Hardware is another area- there’s a renaissance in hardware design and manufacturing. The easy to use 3d design software and printers, powerful COTS embedded computers like Raspberry Pi, and a wealth of turn key manufacturing services, make hardware easier to do than ever before. However, it’s far too easy to underestimate the complexity and cost to bring a quality hardware product to market.
Another potential challenge for both new and older small IoT platform companies is that Amazon, Microsoft, IBM, SAP, Oracle, GE woke up and are offering connected product platforms that promise scale. They have an insane amount of marketing spend to leverage connectivity, device management and applications. While this threat is real and smaller IoT platform companies should take notice, there are still issues with these large company offerings –in many respects they are a “bag of parts” and assembly is required.
I see opportunity in making it easier for a connected product to come to market. CPU core suppliers like ARM and chip providers that license ARM cores are frustrated with the time to market of new connected products, they want higher volumes. Interestingly ARM recognized this issue and acquired Sensinode to deliver a connectivity OS with a back end device management platform, branded as mbed. The premise is that ARM licensees (chip companies) would have an end to end solution to accelerate IoT solutions and resulting sales. A great concept, but it’s not being adopted by the ARM-based chip companies as expected by ARM. Some say standards will solve this issue, but there are too many standards and that is an issue. Making it easier for connected product to come to market is still an opportunity, but not for small companies.
Another opportunity is in gateways, all the major gateway and enterprise router companies are recognizing the market pull for enabling more edge connectivity and processing. I predict the gateways will become a data center at the edge, supporting application containers and a battle ground between ARM and Intel. Let’s say ARM based chips owns the edge device and Intel owns the server, the gateway CPU socket dominance could go either way.
I am also enthused with the rapid evolution of deriving data value with machine learning and trusted data brokering with blockchain.
I haven’t found a pure IOT book yet that I’d necessarily recommend, but there are a couple that I think are outstanding. One is “Hooked: How to Build Habit-Forming Products” by Nir Eya, I find the good and bad of social engineering interesting to understand, avoid the bad, adopt the good. Another interesting read is “The Hard Thing about Hard Things” by Ben Horowitz. A collection of lessons and I knew a lot of the folks. Those are two that might not be on people’s bookshelves that I recommend.