Gartner Recommends Network-on-Chip (NoC) Technology For SoC Design

popularity

In their latest Hype Cycle for Semiconductors and Electronics Technologies report, Gartner Research has taken the bold step of recommending that all enterprises involved in advanced SoC design should seriously evaluate network-on-chip (NoC) technology based interconnect fabric IP:

The technology has continued to receive a good amount of publicity along with continued adoption by leading SoC firms, as a result we have therefore advanced NoC a little further along this year’s Hype Cycle.” — Jim Tully, Vice President and Distinguished Analyst at Gartner.

You can see in Figure 1 below how Network on Chip is on the slope of enlightenment along with through silicon vias (TSV), and is ahead of DDR4 DRAM in adoption.

gartner1

In addition, you can see in Table 1 below a summary table of the hype cycle in the Gartner report which shows how NoC technology provides a high benefit and is less than 2 years from mainstream adoption.

2013-11-20-Gartner Priority Matrix NoC

In the following paragraphs, I will explain a little about Gartner’s Hype Cycle in relation to NoC technology, and also relate the hype cycle to the more familiar Technology Adoption Lifecycle “chasm” popularized by Geoffrey Moore.

NoC Technology’s Position on the Gartner Hype Cycle
In Jim Tully’s semiconductor hype cycle report, he shows that NoC technology will be leaving the “Slope of Enlightenment” to reach the “Plateau of Productivity” soon. (See Figure 1). To support this position he states, “The investments that both ARM and Qualcomm have made in NoC company Arteris further point to increased interest in this technology,” and advised, “All enterprises involved in advanced SoC design should seriously evaluate NoC offerings.”

But what are the “Slope of Enlightenment” and the “Plateau of Productivity”? And what use is the Hype Cycle?

For the purposes of this discussion, I reference Gartner Research’s 2013 report by analyst Jim Tully titled, “Hype Cycle for Semiconductors and Electronics Technologies, 2013”, in addition to Moore’s book, Crossing the Chasm. I also use network-on-chip (NoC) as a technology example and Arteris as a company example because I am most familiar with these and many people in the semiconductor industry are already familiar with these, too.

What is the Gartner Hype Cycle?
Unlike the Technology Adoption Lifecycle or “Crossing the Chasm” model with which you may already be familiar, Gartner created the Hype Cycle to measure and predict expectations regarding new technologies. In other words, Gartner found that for any new technology, peoples’ reactions to it follow the same pattern, no matter whether the technology is the Internet, through silicon via, NoC or even fire and the wheel.

The key here is people. By understanding the reactions of people to new technologies, we can better predict the adoption of new technologies and make smarter product planning (and even personal purchasing) decisions.

The 5 Stages of the Gartner Hype Cycle
Figure 2 below shows technology milestones that occur in each of Gartner’s hype cycle stages.

fig2-gartner-hype-cycle-5-stages-Arteris-noc

 

  1. The Technology Trigger phase is when word of a new technology generates industry interest. Market buzz builds to a loud crescendo.
  2. Then we reach the Peak of Inflated Expectations where visionary companies adopt the technology and push it to its limits.
  3. However, these companies become impatient as they realize the new technology is not a panacea for all their problems. The new technology is characterized as having poor performance, slow industry adoption, and the lack of a clear value proposition. As industry experts and the media emphasize these challenges rather than the opportunities of the technology, we enter the Trough of Disillusionment.
  4. As early adopters gain experience with the new technology, best practices emerge about how to use it. As an example, we can see that NoC technology is on the Slope of Enlightenment.
  5. Then, as the real-world value of the technology becomes accepted by the industry, more companies adopt the technology. A herd mentality starts which results in quick adoption and penetration of the technology. This is the Plateau of Productivity.

A key point to understand is that the time it takes for a technology to travel along the hype cycle is non-linear. In other words, the technology moves through the cycle in fits and starts.

kurt3

To demonstrate this I created Figure 3, which shows how network-on-chip technology has appeared in Gartner’s hype cycle reports in 2007, then 2010 through 2013. You can see how people’s perceptions change non-linearly, with little change between 2010 and 2011, and then a huge increase in awareness and expectations between 2011 to 2013. I do not have data for the years 2008 and 2009, but I’m willing to bet these points are not evenly distributed in time between the 2007 and 2010 data points.

Using the Gartner Hype Cycle and Chasm Model Together
I have seen many people try to relate the Gartner Hype Cycle to Moore’s Technology Adoption Lifecycle and get it wrong.

The issue is that many people simply superimpose the two charts over each other, without realizing that the Gartner Hype Cycle is only valid for the early stages of the Technology Adoption Lifecycle. More specifically, the Hype Cycle tracks perceptions of a technology to a point within the Early Majority phase of the lifecycle, which includes the “chasm” that Geoffrey Moore so eloquently details in his book.

At the left side of the chasm, the technology’s customers and users are “change agents”, looking to get a jump on the competition. But on the right side, the customers and users desire productivity improvements for their current operations. This change in technology users and customers through the chasm is what accounts for the wild swings in expectations shown in the Gartner Hype Cycle!

Accuracy of the Hype Cycle and Chasm Model
How accurately does the Gartner Hype Cycle track a technology relative to the Technology Adoption Lifecycle? As an example, I’ve created Figure 4, which shows the NoC technology hype cycle chart overlaid on the technology adoption lifecycle chart.

fig4-chasm-gartner-hype-combined-2013

From this chart you could say that NoC technology has crossed the chasm and is on its way to fast adoption and hyper growth. If you used Arteris as a proxy for NoC technology, you would have other evidence that supports this. So as a descriptive or backwards looking model, the Gartner Hype Cycle seems to be valid.

Predictive Use of the Hype Cycle and Technology Adoption Lifecycle
What is tougher to answer is whether the hype cycle is useful as a normative or predictive model.

The reality is that there are technologies that start the hype cycle but never cross the chasm and become successful. An example is Floating Body DRAM shown in Figure 1. However, as a technology makes it to the Slope of Enlightenment and the Plateau of Productivity, its probability of market success increases beyond what it was in the earlier hype cycle phases. Therefore, one could feel more comfortable using data from the Slope of Enlightenment to the Plateau of Productivity help make investment or product planning decisions. However, technologies that appear in earlier hype cycle stages should also be researched for future applicability to project plans and investments.

Conclusion
Although the Gartner Hype Cycle and the Technology Adoption Lifecycle / Crossing the Chasm models measure different phenomena, they are related in time and by human beings’ reactions to the introduction of new technologies. The examples used above show how the two models succeed as descriptive models as long as the time frames mapped between them are well understood, and suggest that one could use these models to help make investment or product planning decisions.

Sources
Tully, Jim. “Hype Cycle for Semiconductors and Electronics Technologies, 2013.” Gartner Research report G00252848 30 July 2013.
Tully, Jim. “Hype Cycle for Semiconductors and Electronics Technologies, 2012.” Gartner Research report G00234925 1 August 2012.
Tully, Jim. “Hype Cycle for Semiconductors and Electronics Technologies, 2011.” Gartner Research report G00214777 25 July 2011.
Fenn, Jackie. “Understanding Gartner’s Hype Cycles, 2010.” Gartner Research report G00205030 9 July 2010.
Moore, Geoffrey A. Crossing the Chasm: Marketing and Selling Disruptive Products to Mainstream Customers. Collins, 1991.