Limiters To The Internet Of Things

“Specialization is for insects.” Robert A. Heinlein, Time Enough for Love

In many respects, the Internet of Things (IoT) is already here. But because most of it is not directed at the consumer space, it seamlessly blends into the environment, unnoticed by everyone except for those who are reaping the benefits of it. When we talk about the IoT, most people concentrate on the edge devices, which include one or more sensors, a processor and a means of communicating their data to some higher level router or processing device. That is where we see number such as billions of devices and the size of the numbers makes us pay attention.

The industrial sector is expected to be the largest deployed of IoT edge devices and has long been a user of this technology, although in a slightly different form. Today, IoT is dominated by the consumer space, largely because of the number of phones and tablets that have been deployed. According to ABI Research, “75% of the growth between today and the end of the decade will come from non-hub devices: sensor nodes and accessories.”

Companies, such as Echelon, have been building and deploying devices in this space for over 25 years. The principle difference in that we are migrating from an Intranet of Things to an Internet of Things and that means making devices publicly accessible rather than being private resources. “There is a lot of buzz in the consumer space, with wearables and people with their fitbits, jawbones and nest thermostat,” says Wendy Toth, chief marketing officer for Echelon. “This is all very exciting but is very different from the industrial space. In the Industrial IoT, these are long lived assets that have high value. They are meant to have very limited interaction with human beings once installed and they have some critical requirements related to zero failures and resilience and they must have hardened security.”

Not everyone is convinced that the IoT is ready to go. “We are not yet in the trough of disillusionment,” believes Drew Wingard, chief technology officer for Sonics. “We are still in the upswing, which is the point where you ignore all the failed devices.”

What is clear is that there are a number of limiters to the proliferation of the IoT. The primary limiters appear to be: cost, power delivery and storage, standards, security, and the limits of your imagination.

As costs come down, it will become possibly to extend their reach into more places and additional applications, and the reasons to replace and upgrade legacy devices become more compelling. There has to be a compelling business event or a clear way in which providing the data can be monetized. This is why some still see the IoT as bring in the trough of disillusionment.

“You want to be able to create a design for $500,000 and three months of effort,” says , a Cadence fellow. “This is a highly optimized chip for my application. It is one or two orders of magnitude less from a cost point than what we are seeing today.” The reduction in development costs will be examined in an upcoming article, including changes that can be made in the design and verification flows that could dramatically change the R&D costs associated with creating the devices.

Many devices will be in environments where human interaction is difficult or impossible and this requires very long battery life and thus low-power design, possibly coupled with energy harvesting. The importance of low power design is examined in more detail in the article titled “IoT Brings Low Power To The Forefront.”

Standards enable additional synergies to exist between devices bringing total system development costs down. “Standards will be an important contributor,” says Aveek Sarkar, vice president for ANSYS/Apache. “If there are only a small number of standards on which the ecosystem can operate, then we will see a proliferation of customized devices because it will be easier to develop them. Will an ecosystem develop where a small design team can come up with an idea and will this foster innovation and create more design teams? The bigger design companies will have a lot to say about this, but I expect to see a lot of innovation from small design teams.”

The notion of brown field versus green field makes things different as well. “In the consumer space it is all green field,” says Toth, “but industrial has to take into account what is already in the field. That has to be migrated. There is tremendous opportunity just in upgrading legacy devices. This is easier because standards are already in place.”

One of the ways in which costs are being held in check is that these devices are not chasing after the latest technology nodes. This is not only a cost issue, but also a necessity given the difficulties associated with created analog and mixed-signal devices on small geometries. The larger nodes, such as 130nm and 180nm offer better characteristics for integrating analog circuitry directly, but this only solves part of the problem. The other factor working in the opposite direction is the need for 2.5D or 3D packaging because in most cases it will not be possible to combine the sensors themselves onto the same die. Integrated more of the device on a single die reduces parts costs, reduces fabrication and assembly costs, reduces power consumption, increases security, but not all of the technologies are yet in place to make this a slam dunk.

Security is the primary area that is affected by the migration of edge devices onto a public network. Due to the relentless cost pressure, it is also perhaps the area that is seeing the least development. Implanted medical devices cannot afford to have vulnerabilities that make them open to hacking, denial of service attacks or any of the things that we are becoming used to on the Internet, and this is why Semiconductor Engineering has a whole channel dedicated to the subject.

Imagination Technologies is addressing this challenge by adding hardware virtualization technology into their latest processors, but this assumes that the edge devices will be larger, more general purpose and find other ways to reduce the effective cost of the devices. “Many devices will have multiple tenants offering various services on these devices,” says Amit Rohatgi, president of Prpl Foundation and until recently vice president of strategic marketing for Imagination Technologies. “This is an extension of hardware as a service. For example, a lamp post may detect traffic in the area as well as controlling the light and could theoretically be a point of sales for credit card transactions. This does not require three independent systems talking to three independent back ends, it would be one virtualized system that can handle all of them in a secure and reliable way. The common belief holds that edge devices are specialized but people are starting to push platform as a service because it is general purpose hardware and does not have to be specialized all the time.”

Not everyone sees it this way. Rowen provides the analogy between mammals and insects. “Mammals are fairly general purpose creatures that fill large ecological niches. They are complex and energy intensive creatures and there are not many varieties of them. By contrast, there are many orders of magnitude more species of insects. They are smaller and fit into very narrow ecological niches. IoT devices are like ecological insects. They are highly optimized for a very narrow set of tasks. Each has a unique design associated with the demands of that situation. The average insect has just the features it needs and no extras.”

If the technology and cost hurdles can be overcome, then imagination may be the only limitation. Today, kids are writing apps for smartphones. “Is it going to be the same about ‘chips’ that go into the IoT edge devices,” asks Sarkar? “Probably not, even if the building blocks are all available. There will still be expertise necessary to put the pieces together in the most energy and cost efficient manner.”