This is the post excerpt.


I’m new to blogging as you can tell.  Below you will actually see my first blog on the pharmaceutical industry at the bottom of the page followed by second entry on networking.  I am doing a series of pieces on r&d in American business.  I hope to provide useful background to explain why tax cuts will not produce more r&d.  In fact, government spending (which cannot occur without revenue) is extremely enabling to much of the r&d conducted by American business.  The third in the series will be on social media (about which I know very little but have some thoughts) and software.  The last in the series will be on businesses that rely on r&d but mostly just integrate it into bigger systems.  This includes industries like autos and defense and perhaps I’ll add some thoughts on analytical instrumentation.

The Integration Industries: Automotive and Defense

It used to be that the automobile industry and the aerospace and defense industries developed a lot of new technology.  Today, theose industries have largely become systems integration industries, rather than creators of new technology.  Are there some excceptions to this general rule?  Sure.  GM, for example, has recently acquired a technology firm that is developing a complicated software-driven and sensor-driven package to enable driverless cars.  The general rule, however, is more applicable.  By and large, auto companies do not develop new technologies to make fundamental tax changes in automotive drivetrain systems — the fundamental systems that will enable higher energy efficiently cars.  They do not create new sensors that will enable driverless cars.  They do not create the new electronics and software systems that will fuel new consumer experiences in the automobile.  What they do is apply these technologies and put them together.  In short, they are no longer really more than systems integrators and marketers.

The same is true of the defense and aerospace industries.  Today, those industries seek new partners to manufacture the extremely complicated radar, communications, targeting and command and control systems that are at the heart of defense and aerospace products.

Sometimes defense and automobile companies will acquire a technology star, but they typically do so at a big premium ot the shareholders of these companies.  That means that these industries will not drive huge gains in productivity and make fundamental bets on products that will produce sustainable economic growth over long periods of time to re-establish US economic dominance.

Where does this leave us?  Let’s take a look at the auto industry, which is such an important part of the American economy.  First, the industry will have a lower set of components that produce significant incremental value for the large automobile parts ecosystem.  See http://www.mckinsey.com/~/media/mckinsey/dotcom/client_service/automotive%20and%20assembly/pdfs/the_future_of_the_north_american_automotive_supplier.ashx. Will there be some?  Sure.  The companies like Borg Warner that provide high numbers of gears in with specialized tooling and other investments that make competition very difficult and that enable large improvements in fuel economy will continue to do well.  similarly, companies that have made investments in tooling that enable creation of components with lighter weight materials like aluminum will also continue to do well.  Of course, what will happen with experimental materials that produce even further gains?  I see little happening in the industry to create these new materials or perhaps more impartially mata-materials.

Second, we have reached a plateau of economic productivity in this very large portion of our economy.  See Figure C in the attached.  http://www.epi.org/publication/the-decline-and-resurgence-of-the-u-s-auto-industry/.  And, see also the subsequent analysis in this article on the declining impact of wages in the costs of the auto sector.  That, of course, makes good sense.  After all, as noted above, the industry simply makes less and buys more of what goes into .the cost of the automobiles it makes!  At the same time, the burden associated with workers relating to the cost of health insurance and payroll taxes is exceptionally important to the productivity of this sector — much more so that sectors like networking, pharmaceuticals and biotech and complicated electronics and software that produce much, much more contributed value relative to the cost of their products.

Finally,  look at how much money these entities and some other entities like Comcast and Verizon, which apply technology rather than develop it, spend on advertising.  See http://marketrealist.com/2015/02/automotive-companies-spend-heavily-advertising-rd/.

Let’s look at the public policy implications of all of these factors.

A.  Like the high contributed value industries that I have previously covered — pharma and biotech, telecommunications and networking, software and electronics, these inbdstires will not make fundamental bets on technology to produce long-term sustainable economic development as a result of tax cuts.  They simply are not in the business of making those kinds of bets any more.  They now rely on theose developments produced by others — most likely, venture-backed entities that are funded by venture capitalists who receive their capital mostly from university endowments and pension funds (entities that are not taxable and thus will not increase investments due to corporate or indivudual income taxes).

B.  On the other hand, they will make additional investments in labor and perhaps in sponsoring technology development if the cost of developing new driverless systems and the like if the cost of labor is reduced.  This will happen if there are permanent cuts in the payroll tax and reductions in health care costs.  In fact, note that the CEO of Ford once noted that Ford could benefit substantially from a national health insurance program that actually reduces health care costs.

C.  Unlike the McKinsey report, I believe that the auto industry in the US will benefit substantially from the continued burden of the EPA CAFE standards — the requirements that automobiles meet increasingly tough mileage standards.  The new Asian Tiger economies will be able to integrate automobile components into cars as efficiently as American auto companies.  That’s why we have seen the Korean manufacturers quickly become suppliers of good product as compares with secon-tier manufacturers in the past.  How do we maintain competitive advantage in this climate?  We do so in the way we always have — by creating new technologies that others don’t have.  These include new lightweighting technologies that enable the use of materials that otherwise could not be incorporated into cars.  We do so by productions more complicated transmission systems that also improve mpg.  These include new hybrid and battery technologies.  these also include driverless cars.

The best way to achieve this is probably just to impose a substantial carbon tax, which will drive up the price of gas at hte pump.  For example, a tax of $0 per metric ton would increase the cost of gas by around $.35 per gallon — a little more than 10% increase in gas prices.  This would create a substantial incentive to continue to increase the development of these novel automotive components and I would bet on American business to meet this challenge and increase the competitive position of our auto ecosystem.  These kinds of development activities and new plant investments won’t go to Mexico and Asia.  They will be done here because they need to be close to the source of technology development and because the capital investments needed to implement those technologies have been made here.

Computing and Electronics

The big r&d companies in the computing industry are semiconductor or fabless semiconductor companies.  Their r&d postures are fundamentally different from the r&d postures of the rest of the high tech industry.  Either because the product lifecycle is very short or must meet a standard that holds for a certain period of time, these companies either spend a ton of money on R&D or they might as well just milk what they have in the current generation products and then shut their doors.

Take a look at the numbers in Figure 1 of this link — http://www.electronicspecifier.com/around-the-industry/the-top-10-semiconductor-r-d-spenders-in-2015.  The companies that are trying to sell proprietary chips that are in markets other than low-end consumer markets — Intel, Qualcomm and Broadcom spend well over 20% of sales on r&d.  A company like Qualcomm is always focused on next-generation wireless systems that meet agreed-upon industry specifications.  Its business model is to sell a license to use its technology for royalties and sell chipsets that enable cell phones to work that are designed by them and actually made by an outsourced semiconductor manufacturer like TSMC in Taiwan.  They either bet big on the next-gen product or they are out of the ballgame.  Companies like Intel and Broadcom need to make a series of bets on products that have very short product life cycles.  Like Qualcomm, if they don’t make these bets they are out of business.

Will tax cuts make much of a difference to these companies?  No way.  They either make the r&d bets or they are out of business.


Social Networking and Software

The developments of software until social networking was a huge contribution to economic productivity.  The digitalization of analog functions has led to large increases in productivity in manufacturing, instrumentation as well as all of the sales, marketing and accounting functions of business.  In fact, these software developments and the great increase in Internet applications in the 90’s fueled huge increases in US productivity.  See http://www.frbsf.org/economic-research/publications/economic-letter/2015/february/economic-growth-information-technology-factor-productivity/.   Especially if you look at Table 3 of the study focused on various sector’s contributions to productivity, you will see that the 90’s saw the largest increase in productivity in the IT industry itself.  That increase in productivity has ongoing impacts on productivity especially in It-intensive industries like banking, insurance and health care.

We cannot expect that kind of productivity growth to occur very often in our lifetimes.  That’s another reason why the Trump Regime projections of 3% or more economic growth are beyond folly.  Look at a simplified version of the history of the computing industry.  We went from the large computer systems with “dumb” terminals (display at the desktop and all processing done by the central computer) to client/server networking (lots of processing distributed to servers outside the core as well as processing at the desktop personal computer) to Internet and then cloud based systems in which portions or all computing is returned to centralized computers run by entities outside of the enterprise.  There was one time when instrumentation and manufacturing systems were computerized.  That occurred mostly at the end of the 20th century.  The increases in productivity by moving to the Cloud will have some effect, but nowhere near the effect of moving from an analog control system to a digital system.  Similarly, the move to the cloud from Internet based systems will continue to increase productivity but not as much as the move from client/server networks to Internet-based systems.  That’s why we saw the HUGE increases in productivity in the 90s and will not see that again for a long time.

There’s another reason why software spending will not produce the kinds of productivity gains like the gains we’ve seen in the past.  That reason is the incredible focus on social networking.  To understand that, it’s important to understand what drives social networking and it is not being social.  To the contrary.  Writing this blog versus actually talking to other human beings is the worst excuse for being social of anything I can imagine!  What drives social networking is advertising revnenues.  Take a look at the numbers.  See https://www.recode.net/2017/5/2/15516674/global-ad-spending-charts.  I used to think that Google and Facebook were just TV channels.  I was wrong.  Google and Facebook are dwarfing Comcast (NBC), Disney (ABC) and CBS.  In fact, the combined ad revenues of NBC, ABC and CBS are only about equal to the ad revenues of Facebook, and Google’d ad revenues are about 3x the revenues of Facebook!  So, Google and Facebook are media channels just like TV networks.  The big difference is that they don’t produce their own content (well, the TV networks don’t really produce all that much either).

So, do Google and Facebook and Twitter and Instagram, etc. for that matter increase productivity.  Sure, but not very much.  Consider the best search engine before Google — Alta Vista.  The truth of the matter is that if you could learn to use a few Boolean Algegra methods (you didn’t even need to know Boolean Algebra or why it worked; you just needed to learn how to use them) Alta Vista worked almost as well as Google and I’m sure if there were no Google Alta Vista would have continued to get better.  Who knows?  It might have even gotten to be better than Google.  The point is that Google has contributed to productivity with its search engine, but it’s hard to say how much.  Clearly, it has also contributed to productivity with some other tools like Android, Crhome and even in the cost per bit of its information handling in its data centers, but how much?  Clearly, it’s all more about making the advertising revenue engine be more efficient.  And, why shouldn’t it be?  That’s what Google is in business to achieve — lots of profits for its shareholders.

My point is that the shift in so much in software and computing resources to social networking limits the amount of productivity benefit we get from this kind of software spending versus the software spending that took analog functions and made them digital.  Consider looking at things another way.  A lot of what Google and Facebook achieve for their customers (not their users who are not their customers but the advertisers who are) is to create winners versus losers.  And, at some point it makes no difference to productivity if distribution channel A wins by getting its ads in a better place on Google than distribution channel B.  Will the economy as a whole do better if the Gap wins the digital ad wars versus The Limited?  For those who care more about power cleans than clothes, will it increase the productivity of the economy if Rogue Fitness does better than Pendlay?  (That last sentence is for all my old CrossFit friends.).

I’m not saying that Google and Facebook do not create technology.  They do — especially Google, which has done some pretty amazing things.  What I’m saying is that you just cannot compare what they are doing for the world with going from the IBM mag card to the Internet or from no microprocessor in an instrument to the programmable logic control board or to the FPGA chip.

Here’s a simple example.  The first millimeter wave radios made in very low quantities cost $75,000.  Volume brought the cost down a lot, but what will really bring the price down is converting the integrated circuit in the radio from a compound semiconductor to a CMOS chip with volumes of hundreds of millions or billions of units.  When that happens, the chip will cost way less than a dollar and the cost of the systems will be about 20x less than the cost of the original systems.  That kind of thing can have a massive impact on economic productivity.  Spending money on optimizing advertising algorithms cannot come anywhere close to that kind of productivity gain.

Can there be software developments that produce large productivity gains in the future.  I’m sure there can be.  With a small team, I produced an automated and patented system for extracting information about mutual funds from the SEC’s EDGAR database.  That produced some productivity gains for a small sector of the American economy (we thought it would be much bigger when we developed the algorithm).  I’m absolutely certain that there are many segments of the economy that will still benefit from cloud computing, new data distribution channels, new processing methods, new algorithms and the like.  I just believe that the changes in this sector will be more incremental than revolutionary going forward.

Moreover, I believe that the best way of developing these novel ideas is in venture-backed organiazations, not in big and profitable companies.  By definition, only the big and profitable companies can benefit from tax cuts.

So, let’s get back to the original reason that I’m writing these blog posts.  A tax cut for American business or rich people is not going to produce sustaninable economic growth.  The reason is that the industries that have produced amazing productivity gains will not produce the kind of productivity gains in the future and whatever productivity gains will arise will probably largely come from new venture backed companies and not from the big profitable companies.  After all, Google and Facebook — like their pharma counterparts — have become major acuqirors.

More on Networking

If you follow either Arista or Cisco or the stock market in general, you may have noticed that Arista won two matters before the Patent Appeals Board — invalidating two patents on which Cisco sued Arista for patent infringement.  Neither patent was of great significance in any event and this is what we’ve come down to.  After years of reaping the benefits of packet switching with a business model that enabled high profits by requiring purchase of Cisco hardware to get access to its proprietary software, Cisco is now subject to competition enabled by software defined networking.  Arista and home brew do it yourself systems have used software defined networking to compete effectively with Cisco.

Arista has done a great job of taking away market share from Cisco, but I wouldn’t bet long term on either Cisco or Arista because both companies are still betting on packet networks.  As I explained in my previous post, the hype about efficiency of packet networking is overblow because of the need for overprovisioning.  Similarly, the hype about reliability of packet networks has not produced telco-grade reliability  of ISP’s adn major websites.

Someone is going to make a fundamental bet on an optical circuit switched network to overcome these barriers and the prior barriers to circuit switching.  The fundamental issue is to increase the number of wavelengths (colors) that can be transmitted in optical fiber.  The current limit is based on the need to maintain coherence without chromatic dispersion.  In other words, the colors cannot interfere with one another.  Once that nut is cracked, it will not be difficult to create a circuit switched network with a top-level allocation algorithm, a frequency generator that is like an atomic clock and the interferometry system needed to groom the wavelengths.  This would result in a massive number amount of bandwidth for each computing node — the kind of efficiency that would surpass the efficiency of any packet network.  And, reliability of the network should be pristine because the amount of bandwidth per computing node would make the system effectively a permission-based networking system.

Cisco, Arista, Juniper and other big companies have nothing going on in this regard.  For that matter, I don’t know of anyone else trying to do this. It only is likely to be pursued by a physicist who has the smarts to acquire a great engineering team and the ability to communicate the idea to vc’s with enough financial backing and guts to pull this project off.  That will be a bet on the future that will create sustainable economic growth.  It will not occur because of tax cuts.  It will occur because of a combination of vision, will and capital that is not constrained by the need to produce quarterly results to satisfy Wall Street expectations.  Donald Trump and the Republican congress don’t seem to understand that all.

Trump’s Abandonment of Paris

I will interrupt my blog posts on R&D to pay attention to Trump and the environment.

Trump’s abandonment of the Paris Treaty is, of course, based entirely on lies. It is not true that energy prices are increasing. They are in fact historically low. For example, look at hte real prices of electricity for the 51 years from 1960 – 2011. https://www.eia.gov/totalenergy/data/annual/showtext.php…. They’re substantially lower. How about the price of oil? Sure, it was high for a good chunk of time, but they are way down starting in 2015 and remain very low today.  https://inflationdata.com/…/Inflation_Rate/Historical_Oil_P…. Moreover, they are projected to stay low for a very long time. Look at the amazing stability of oil and natural gas pricing — even with the Obama Clean Power Plan in place. https://www.eia.gov/outlooks/aeo/data/browser/….

Are we losing jobs due to environmental regulation? Nonsense. Consider that the US auto industry has thrived in the Obama years with very tough CAFE standards in place. Consider that companies like Borg Warner, which manufactures more complicated automatic transmissions — especially for hybrid and electric vehicles — and turbochargers. This is a company that actually is thriving based on the demands of additional regulation.

Look at the passage of the Clean Air Act of 1970. One of the great pieces of history of my former employer — Thermo Electron (now Thermo Fisher Scientific) was that the VP of R&D of Ford Motor came to Waltham to review the status of the company’s steam engine project for Ford (not a success). During his review, however, he met with George Hatsopoulos (CEO) and complained of the inability of Beckman in making progress to produce an instrument to measure oxides of nitrogen to levels that would be required under the new Clean Air Act. Hatsopoulos then took the Ford executive downstairs to look at the prototype chemiluminescent detector that they were using to measure oxides of nitrogen on the steam engine project. The Ford executive said to Hatsopoulos bring a version of the instrument to us at our Michigan R&D center in 90 days and we’ll give you a purchase order. Sure enough, the proeudcut was ready — although that meant an all-night drive by a Thermo engineer to get the product there in time. Thermo would then set the standard for NOx detention and later SO2 detection with pulsed fluorescence technology and a slew of other instruments. Today, Thermo is the world’ largest analytical instrumentation supplier and has a market capitalization of $68.4 BB.

How about renewable energy jobs? They are growng 12x faster than the rest of the economy.  http://fortune.com/2017/01/27/solar-wind-renewable-jobs/

How about Trump’s lie that business wants what he’s selling. Total nonsense. The new =CEO of Exxon Mobil says no. Elon Musk says no. A recent survey of international business people say that they are ready for carbon pricing. See http://www.ey.com/…/ey-shifting-the-carbon-pricing-debate-n….

Folks, we cannot afford to wait for 2020. We need to act on climate now. MA has a carbon fee proposal in the legislature now (actually, two such proposals, but the revenue neutral version S. 1821 probably has the only chance of success). We must adopt carbon pricing in any jurisdiction that has any chance of its adoption. If we can get to critical mass among the sates in adopting carbon pricing — perhaaps about half the population of the country (CA already has cap and trade, which represents about 13% of the US population), we can ameliorate much of the harm that Trump and Pruitt will do to the environment. We must do all we can to support these legislative efforts. In Massachusetts, about 1/3 of the legislature has signed onto S. 1821 as co-sponsors. We need more and we need to convince both the leadership and the governor not to stand in the way. Nothing short of the future of the world is at stake here.

Telecommunications and Internet Networking: Limits on Optimality


In my first post, I discussed the limits on sustainable pharmaceutical development.  The state of the industry does create an opportunity for a dynamic biotech industry as well as many opportunities for the venture capitalists that support them.  Ultimately, optimality is not achieved though because the short-term focus on earnings and the overall culture of pharmaceutical companies limits the ability of the pharmaceutical industry to make big bets on the future that will propel the kind of sustained growth we saw from the pharmaceutical industry in the postwar period is not likely to be replicated.  No amount of Trump tax cuts will change that.  In fact, if those tax cuts cause a reduction in NIH funding of basic research, it is likely that the pharmaceutical and biotech industries will suffer over the long run.

In this post, I now focus on limits on sustainable growth in telecommunications and Internet networking, an industry that produced massive economic growth in the last decade of the 20th Century and the first decade of the 21st Century.  The poster child for that growth is Cisco, which has reached almost $50 BB in sales.  Unfortunately for both Cisco and the US economy, its growth has now abated and it has made severe job cuts.

The engine that propelled Cisco’s growth — the router and the accompanying IP protocol — have also been the basis for the development of the Internet.  Let’s return to Al Gore’s claim that he was crucial to the creation of the Internet (No.  He never said that he “created” the Intenet).  Well, Gore did not play a direct role, but Federal funding surely did.  The original IP Protocol software and the related development of he router occurred in research laboratories in Massachusetts and California.  In Massachusetts, the original ARPANET network nodes and routers were deployed by Bolt Beranek & Newman.  On the West Coast, similar research occurred at Stanford and XeroxPARC.  This work was important because it allowed for the development of the IP protocol, which enabled the transfer of data among disparate kinds of networks regardless of protocols used within those networks.  The data between the networks was chopped into packets and sent from one node to another with the router — powerful computers that transferred the packets from node to node electronically.

Needless to say, the idea took off in the 90s and powered the development of the Internet (Cisco did acquire and develop other technologies, but the router proved to be its key growth engine).  Of course, an interesting twist of fate is that the original Cisco software program that powered its routers was the property of Stanford, which complained to authorities and then criminal charges were pursued against Cisco’s founders.  The case was dropped, however, when the parties came to an agreement that allowed Stanford to share in the Cisco bounty.  the bounty proved to be substantial.  Cisco began its march up the so-called S curve in the 90’s and sustained that growth until the last five years.  From 2015 to 2016, its revenues were essentially flat after having lackluster growth for the prior few years.  Last quarter it actually had a decline in revenues and announced that it would cut another 1100 jobs after much larger layoffs last year at Cisco and much of the tech industry.  See http://fortune.com/2016/08/18/expect-more-tech-layoffs/.

A couple of factors account for the Cisco decline — both relating to software defined networking technology developed to a great extent at Stanford.  The technology is open source, though, so Cisco could not dominate the market.  Using SDN technology and the open source Linux operating system, former Cisco engineers developed a new company called Arista that developed their own routers to compete with Cisco.  While Cisco was leveling off, Arista grew from $193 MM in sales in 2012 to $1.13 BB in sales in 2016 — a 42% compounded growth rate.  The new software allowed users to develop their own networking protocols with the Linux software program and with th common platform for all Arista routers.  In addition, many users started to develop their own network nodes using off-the-shelf routers and the open source SDN software (as com[pared wit h the proprietary Cisco software).

Those are significant developments for Cisco in the short run, but I believe that the entire networking industry faces a greater challenge and one that very few — if any — companies are addressing.  The problem lies with packet networking as a whole.

Packet Vs. Circuit Switching

Let’s consider some of the claimed benefits of packet switching.  The first and most obvious is that it’s more efficient.  Here’s the idea.  If you have a circuit-based network (Baby Boomers — think back to the Lily Tomlin operator routine on Laugh-In), you have to complete the circuit for the entirety of the information you want to transmit.  If you have to reserve the circuit for the entirety of the information transformation period, that seems like you need to have too much downtime of the network to enable reliable transmission.  So, it seems like packets would be the way to go — especially if you are transmitting data rather than voice phone calls because distributing data like videos require much more bandwidth.  There’s a fly in the ointment, however.  It turns out that packet networks have to reserve a lot of capacity as well.  There are lots of reasons for that.  Packet networks can get congested and the points of congestion lead to potential delays and we all know that when we click on something the last thing we want to see is the hourglass or the swirling ball.  We want our download when we click or at least soon thereafter.  This delay is called latency.  There are other problems that actually can be more significant.  For example, because the data is divided into packets the delay can result in losing the data altogether when there is too much congestion.  And, another factor is that internet service providers simply have to have some extra capacity for the bursts in data that arise from time to time.  As a result, data centers use only 12-18% of their capacity!  See Section 2.2 of the following study prepared for the Natural Resources Defense Council.  https://www.nrdc.org/sites/default/files/data-center-efficiency-assessment-IP.pdf.

Another claim of packet networks is that they are more robust or reliable than circuit networks.  The gold standard in the telecommunications market is that the network operate on a 5 9’s basis.  That means that it’s up and running 99.999% of the time.  In a year, there are 24 hours in a day, 60 minutes in an hour and 365 days in a year.  That totals 525,600 minutes in a year.  At five nines, that means that they are up and running almost all of the time — leaving downtime room of only 5.26 minutes a year.  Whoa!  If you have an Internet Service Provider that is down only 5.26 minutes a year, please tell me what it is.  I’m all in for that.  Look at Table II in the following study.  http://iwgcr.org/wp-content/uploads/2014/03/downtime-statistics-current-1.3.pdf.  YouTube is the only website that is up on a 5 9’s basis.  Here’s the data on the big cloud service providers.  http://www.networkworld.com/article/2866950/cloud-computing/which-cloud-providers-had-the-best-uptime-last-year.html.  Amazon Web Services is the best at 99.972% uptime, which is really good.  Most others are not so good.

There is nothing inherently unreliable about circuit networks.  It’s not clear that you can say the same thing about packet networks.  WE can at least understand that packet networks may be as unreliable or perhaps more unreliable than circuit markets.  First, packet networks send the routing information in-band (i.e., with the data to be communicated).  If the data is held up due to congestion, so is the routing information and the congestion may result in loss.  Second, the routing algorithm in an IP network can be very complex, which can result in either congestion or misconfiguration.  The uptime data seems to confirm the notion that packet networks have an inherent level of unreliability (at least if you want to maintain control over costs.

Optical Switches

So, why isn’t there more optical-based circuit switching?

Let’s delve into one more piece of background information.  In the late nineties and the turn of the Century, the technology industry perceived the inexorable growth of the Internet.  As much as we have seen phenomenal Internet growth, the Net has not grown as fast as predicted back then.  As a result, both telecommunications experts and the financial community thought that the telco network would have to include new all-optical elements in the form of optical switches.  Surely, according to the consensus at the time, the only way to keep up with the inexorable growth of the Internet would be to use optical switches that don’t have to be swapped out any time you increase the line rate — the speed at which data is transmitted through the network.

Two factors got in the way of optical switch adoption.  One was that Internet data was “only” growing at an approximate rate of doubling every year rather than every six months.  The other factor is that the development of optical switches did not move along as fast as projected.  That led to a massive expectation/achievement gap in the technology and financial markets as ~$5 BB was invested by venture capitalists and public companies like Nortel and Corning in optical switching — all of which ended up as pretty worthless.  The switches that were produced did not meet expected specifications with respect to size, cost, flexibility, etc.  Moreover, these switches did not do what packet switches do — “groom” data traffic so that it could end up at the right network node.

There is one kind of optical switch, which actually is based on a hybrid technology, that did get significant — but not huge — adoption in the telecommunications network, the ROADM — reconfigurable add drop multiplexer.  This device allowed large flows to be passed through the optical part of the switch and allowed traffic to be added or dropped at network nodes along the way with filters or other methods.

In the meantime, the router market continued to grow at a fast rate to keep up with Internet traffic and with the significant growth in data centers supporting social networking, distribution of video and music and large amounts of corporate data.  Cisco does a really good job of tracking this data and publishing it.  Here’s their data for the period from 2015 to 2020.  http://www.cisco.com/c/dam/en/us/solutions/collateral/service-provider/global-cloud-index-gci/white-paper-c11-738085.pdf.  Look in particular in table 1 of that piece.  Data center traffic in the cloud is massive and is growing at a compound growth rate of almost 30%!  That’s a massive growth rate.

Now, can we overcome the problems with optical switches to see them adopted in networking applications.  There are some limited uses of them today in addition to ROADMs.  Look at the data on how much of the traffic stays within the data center.  Almost 80% of data center traffic stays within the data center for the large hyper scale data centers that number is more than 80%.  Consider what this means.  For every piece of data that goes to the cloud, 4 times the amount of data moves within the data center.  You just don’t need the granularity of the router to move that data around in most cases within the data center.  Unfortunately and this is where we get into the public policy implications of this, most of the big data center owners and operators do not use optical switching to move very much data around.  Instead, they use the same routers that we have been talking about.

The implications are high cost (big routers are very expensive — especially proprietary ones, which can cost almost $1 MM).  They also take up a lot of space and space comes at a premium in data centers.  Finally, routers are the biggest energy hogs in data centers.  A Cisco core router can require more than 9,600 Watts versus an optical switch, which may require as little as 35 Watts or as much as 85 Watts — two orders of magnitude less.

Several of the big data center owners and operators recognize that they eventually need to adopt optical switching in the data center to send large elephant flows along relatively stable paths within the data center.  Only one of the major data center owners and operators has published that it is following this path.

In addition to the data center, some academics have proposed creating hybrid networks that combine packet and circuit switching.  One of the leading advocates of this is Professor Nick McKeown of Stanford.  See http://yuba.stanford.edu/~nickm/papers/191340-2a.pdf.  Some telecommunications companies recognize that like data centers they can increasingly adopt software defined networking in their core networks.  None of them have sought to create these hybrid networks.

These represent incremental changes.  SDN represents a significant advance, but it does not fundamentally change the nature of networking.

Can there be radical changes that eliminate the router altogether.  I participated in planning related to this kind of network.  While we received some conceptual funding from one government agency, we never received enough funding to get the idea off the ground, although we did obtain a patent on the idea.  I suspect it would take so long to get the idea off the ground in the current climate that the patent will never have any value.  The basic idea is that you can eliminate the router if you can create a way of using more than 40 wavelengths (frequencies or colors) in fiber optic cable — substantially more.  That would require sub-wavelength resolution and we contemplated a method of achieving this.  I fundamentally believe it can be done with an investment of perhaps $25 MM.  The rest of the network would require some new algorithms for assigning colors and a frequency generator, optical switches and some very high speed modems at or close to user devices because computers today could not receive data at the rates I can imagine in this kind of network.  The frequency generator would be similar to atomic clocks that already exist.  The algorithms would require some optimization, but should not be difficult.  All told, you could probably bring this networking idea to fruition for something like $50 MM.  We calculated that this networking method could cost-effectively deliver one gigabit per second service to every computer in the U.S. with high reliability, quality of service and probably greater security because circuit switching could probably function as permission-based information distribution.

That is a trivial investment for a system that would revolutionize networking and propel both commercial success and job generation for a long period of time — perhaps longer than the nearly three decades of Cisco’s dominance.  Will it happen?  Perhaps some day, but why not now?  Public companies won’t do it because of the constraints on R&D spending and the lack of will caused by their inertial cultures.  It’s possible that a company like Google would do it if it were motivated.  Google certainly is willing to make big bets on infrastructure designed to give Google a cost per bit advantage over any of its competitors.  Venture capitalists are not likely to pursue this kind of idea because at the end of the day they are also somewhat conservative and would fear that Cisco would crush this nascent effort so that it can maintain its router franchise even though it’s starting to lose that franchise today.

The key point for me, however, is that tax cuts have absolutely nothing to do with the implementation of this idea.  Saving a little bit of money in a profitable entity is not the make or break issue for this kind of idea.  A new generation of networking technology may come out of an academic effort, a government-funded effort or a venture capital effort.  It is highly unlikely to come from any other source.  Tax cuts will not help the entities that might consider adopting a new network paradigm.  

My First Post

Hi Folks,

I have been posting on public policy issues for the last couple of weeks on Facebook and have decided to put my posts on public policy issues on this blog.  I had started to post on how US business has failed to make fundamental bets on the future through R&D.  My point is that no matter what the government does to create incentives sustainable business development is limited.  The reasons relate to the perceived need of public companies to meet the short-term earnings objectives tracked by Wall Street analysts and the overall business culture that pervades these companies as a result.  I’m repeating the first of these posts from Facebook here and look forward to additional posts from time to time.

What’s wrong with American business? My point is that no matter what the Federal Government does to incentivize business we still have some fundamental barriers to the kind of sustainable economic development we need. I believe that the problems relate to two factors. They are both a function of the first probem, which is that publicly held companies are so focused on short-term profits due to the demands of Wall Street that they do not make long-term bets on r&d. The other issue is that this short-term focus has led to an ecosystem within and without the companies that reinforces the short-term focus. Of course, the US tax system contributes a little bit to this as well because advertising and marketing expenses are by and large (the r&d tax credit being the exception) treated the same as r&d expenses. To a US CEO, why not go after short-term gain when both the public markets and the tax system reward you for doing so?

To peel back the onion further on this idea, consider the big pockets of r&d in this country. Let’s look at the industries that spend the most on R&D. The big spenders give us the guideline to see what we’re talking about. See https://www.strategyand.pwc.com/innovation1000. The big spenders are the pharamaceutical industry, computers and electronics, the auto industry and telecommunications and Intenet networking. We can also add defense electronics to our analysis, which has a lot of r&d but it’s mostly funded by the Government. Let’s do this one post at a time — the first on the pharmaceutical industry.

The Pharmaceutical Industry

Consider that the pharmaceutical industry used to be one of the darlings of Wall Street. The reason was that it was amazingly profitable (80-90% gross margins and outstanding bottom line profits and returns on equity) with stable growth and huge amounts of cash. In the postwar period, the drug companies built these amazing financial engines because there were a lot of fertile targets and the pharmaceutical industry used their huge libraries of small molecule compounds to find new drug targets by rote. Then pharmaceutical chemists could optimize the druges for development. After many years of clinical trials, results when good enough would produce new drugs with a big enough pipeline to assure long-term profitability.

Eventually, the targets reduced and the nature of drug development was fielded ny new science ushering in the new biotech industry. The first generation of biotech druges sought to mimic naturally occuring proteins that would have desired therapeutic value when injected into the body. That only lasted for so long. There was an effort to use biological tools to create so-called rational drug design but Mother Nature proved too complicated. For awhile the biotech industry — like pharma — produced lots of dry holes. Then a new generation of antibodies plus small molecules that were found using biotech tools followed by immunotherapeutics and even some limited success with gene therapy led to a new wave of biotech drugs.

Throughout this whole period, the pharma indsutry did produce some drugs — just not like the huge pipeline of drugs that were commercialized during the few decades after WW II. Left without this pipeline, they made Wall Street modestly happy by buying or doing partnership deals with the biotech industry. There is, nevertheless a long-term shortage of capital for big bets on the health care future. See http://www.nature.com/nbt/journal/v29/n1/full/nbt.1748.html.

What’s wrong with the current situation — the biotech industry providing a huge chunk of the innovation for drug development as a whole? The simple answer is that it would be no impact at all if there were enough capital to go around among the pharma industry to buy out the biotechs. The problem is that there is not. First, many biotech drugs never get far enough down hte pipeline to attract big pharma to buy them out. They become fallen angels. Another problem is that the pharma industry is not generating enough long-term gains on its own to support the market caps needed to buy out the biotechs. The biotech in the meantime mostly rely on biotech buyouts because they cannot get sufficident capital to become the next Amgen or even Biogen.

There is a good long-term solution. The pharma industry needs to start looking much more like the biotech industry. It needs to convince Wall Street that it makes sense to make big bets on the future and sacrifice some of their current earnings. They neeed to restructure their companies so that there is an ecosystem supporting the development of big bets on science. In that way, there will be winners and losers but the winners will be huge winners who become engines of ongoing drug development that makes big pharma look a lot like Genentech.

Will it happen? I doubt it. Wall Street and existing company ecosystems are very difficult barriers to overcome. It’s not just the pharmaceutical industry that needs it. Our country desperately needs industries that look like the phrama industry of the sixties and seventies.

The next post will be on telecommunications and Intenet networking.

The 12th edition of Strategy&’s Global Innovation 1000 Study focuses on the mix of R&D spend and the shift from physical product-based solutions towards software…