中文 English

Hyperscaling The 21st Century Engineer

A lot has changed technologically in the last 20 years, but underlying themes hold steady.

popularity

While January is the month of predictions for many, I have made it a habit to look back and see how previous forward-looking assessments have worked out. It is fascinating to see how many past predictions were off and how little has changed in some areas.

Twenty years ago, in January 2001, the front cover of IEEE Spectrum set the theme of ubiquitous connectivity in an always-on world. Some of the predictions in “Welcome to the Always-On World” by Philip E. Agre in his time at UCLA were scarily accurate. Agre writes that:

“increasingly freed from geographic constraints and equipped with powerful search tools, we will be able to pick out exactly the people we want to associate with, and we will be able to associate with them whenever we want. We won’t devolve into disembodied brains, of course, and geographic proximity will always play an important role in our lives. The point, rather, is that we can maintain more continual relationships with whomever we associate with, near or far.”

Agre predicts a future of email being checked anywhere, anytime, which has been the reality for quite some time. As indicated, security will be critical in the relevant industry verticals of consumer, mobile and networking.

In regards to the industry vertical of transportation, Elisabeth A. Bretz’s “The car: just a Web browser with tires” gives a great outlook. “Why not think of everything as just an internet node,” Scott McNealy is quoted saying, “so a car is just a [web] browser with tires.” Today, 20 years on, we are getting there, with pretty much full connectivity. At the time, the article focused a lot on “component makers working overtime to develop Bluetooth-enabled cars.” What is increasingly worthy of further attention in this day and age is the interoperability of old and new “devices”, i.e., cars, on the same network of streets, in what some call “long lifecycle markets. A full quarter-century ago, in January 1996, the transportation section of IEEE Spectrum focused on how satellites will supersede microwave landing systems, how faster, smarter ships shimmer on the horizon, how airplanes get help against microbursts and how—at the time—Texas was leading with intelligent highways.

The same issue in 1996 features an article about the aerospace and defense industry that talks a lot about the cost of development. The considerations regarding competitiveness in the viewpoint “Getting ahead of the pack—and staying ahead” by Don Fuqua, at the time president of the Aerospace Industries Association, still feels modern even today. Some things stay the same. Fuqua also says that “after a decade or so of a dedicated quest for new efficiencies, many companies have mined the traditional approaches to cost-cutting and productivity improvement for just about all they can yield. Taking competitiveness to the next level demands approaches like ‘flexible manufacturing’.” He later defines this as “expanding computer-aided design and manufacturing (CAD/CAM) to include making a part or component from scratch by computer-controlled tooling. In a broader sense, it gives competitive actors—cost, quality, reliability and less development time—the same status among design considerations as a product’s performance.”

Last year, Will Roper, the Assistant Secretary of the Air Force for Acquisition, Technology and Logistics, introduced a blueprint for “Disruptive Agility for a Disruptive World,” similarly focused on cost, avoiding re-spins and the application. Also, Augustine’s laws came out in 1986, stating that “in the year 2054, the entire defense budget will purchase just one tactical aircraft. This aircraft will have to be shared by the Air Force and Navy three and a half days each per week except for leap year when it will be made available to the Marines for the extra day.” So, it is fair to say that the topic of cost was, is today and will be for the foreseeable future, the main topic to be addressed in this industry vertical.

So, what does all this mean for us engineers? We need to hyperscale for the era of hyperscale computing!

In the January 2001 issue of IEEE Spectrum, Joseph Bordogna’s article “The 21st Century Engineer” identifies five capabilities that are shaping the future of engineering: terascale, nanoscale, complexity, cognition, and holism.

Terascale would take us “three orders of magnitude beyond present general-purpose and generally accessible computing capabilities.” 2001 would see the first Apple store opened, the release of MAX OS X and Windows XP, the first Microsoft XBOX and iTunes’ first release (Source: Computer History Museum). Twenty years on, we have cloud-based scalable availability of compute.

Nanoscale “will take us three orders of magnitude below the size of most of today’s human-made devices.” Okay, no question there. Ten years ago, I wrote about us “discussing 16nm/15nm technology nodes.” Today we are discussing 3nm and beyond, as well as “More than Moore” techniques that allow us to continue to grow in complexity. Speaking of which, Bordogna’s article referred to complexity to the point “where the components of a system never quite lock into place, and yet never quite dissolve into turbulence, either.”

Holism is “the concept that an entity is greater than merely the sum of its parts.” Bordogna concludes that the “hallmark of the modern engineer is the ability to see connections among seemingly disparate components, and to integrate them in ways that exceed the sum of their respective capacities.” One could argue that while disciplines are increasingly connected—software and hardware, electrical and mechanical, and so on, engineers who can comprehend and become experts in all fields are still rare. Instead, communication and interfaces between tools and, of course, the use of abstraction, has become central and will become even more so going forward.

While the absolute numbers that we reached in the past two decades may be debatable, directionally, all these considerations were remarkably accurate. “Scale” itself is the key, and it will be interesting to see its continuation in a hyperscaled world. And as to changing requirements, given how some of the challenges were so imminent for decades already, William Gibson’s 2003 quote seems to apply today more than ever: “The future is already here—it’s just not evenly distributed!”



Leave a Reply


(Note: This name will be displayed publicly)