Modern Era Networking

How do you know you are talking to a person who knows very little about modern networking? When they tell you that they think next-generation networking is about building a spine/leaf networks with legacy protocols so they can have some OS portability like the webscale companies. If you have been reading this blog for five years, I would first like to say thank you and second I am certain that you know the answer is not reinventing the past. I have been writing about the emerging Modern Era of networking for a few years. Continue reading

Four Years Later…

Today is my four-year anniversary at Plexxi. I was in New York the week before Christmas to attend an investor conference focused on security and networking. It was a two-day trip that I expected to go by quickly as it was full of meetings and dinners. A colleague and I met with a number of crossover investors, analysts as well colleagues in compatriot companies. In our very first meeting an investor asked “four years in, how has it turned out compared to how you thought it would go when you started?”

Continue reading

Vignettes I Heard this Week….

Another week in a startup, means another week on the road.  I was in BOS, NYC, SFO and Palo Alto this week for a mix of customer meetings, conferences and networking.  I enjoy spending time in PA because I always get a good mix of topical conversation with VCs, entrepreneurs, colleagues, customers and others.  Here are some things I overheard that made me think:
Continue reading

Nicira, SDN and Travels

It has been seventeen days since my last blog post, which is one of the longest time periods between posts, all due to summer holiday, busy work schedule, Board meetings and family visitors.  A lot has happened in the intervening few weeks and here is my take:

Nicira was acquired. Several hundred people asked me what I thought of the deal and a subset of this group asked me why I did not blog about the deal.  I think most people know I work at Plexxi and I think most people know that Nicira and Plexxi share a common investor in Lightspeed Ventures.  My first reaction is congratulations to the Nicira team; congratulations to John Vrionis and the LightSpeed team.  I read a lot of analysis of the deal and I thought this one was interesting.

I also thought this summary of the deal significance was pretty good.  I do not know the author, but from the preciseness of the numbers I wonder if he has a Nicira presentation.  From the post “Nicira claims that its solutions save between $8.8 million and $17.6 million in reduction of server over provisioning and $8 million to $16 million in reduction of cost per port in a 40,000 server environment.”

I always viewed Nicira as a complement to what we are doing at Plexxi, but outside of their material on their web page and what I glean from common customers and industry colleagues, I have never seen a Nicira customer presentation.  I would enjoy reviewing one, but the web seems to be void of materials.  That might seem puzzling as we share an investor, but typically portfolio managers do cross pollinate information between investors.  If anyone can point me to presentation, please let me know.

SDN Stuff: Serendipity does occur.  Four weeks ago I recorded a podcast with the Packet Pushers team on whether SDN is a game changer.  I was asked to participate independent of my affiliation with Plexxi.  Here is a link to the podcast on the Packet Pushers site.  Ethan and Greg look like geniuses because we recorded this podcast on July 1.  Four weeks later in light of the Nicira acquisition there are $1.2B reasons to think that SDN might be disruptive,  Finally, a thank you from me to the Packet Pushers team for their gracious invite.

I am off for a week in SJC/SFO with numerous customer presentations and Plexxi system trials.  In the course of this coming week I know I will be seeing a number of readers and customers in the SDN ecosystem.  As of Sunday, I still have a few breakfast, coffee and beer slots open; email or DM if you want to meet.


* It is all about the network stupid, because it is all about compute. *

** Comments are always welcome in the comments section or in private. ** 

How Arista Networks affects Venture Capitalists

Back in May I wrote this post on the deficient state of venture capital funding for large scale networking projects.  At the time, I was not thinking about green tech and companies like Solyndra.  I was focused on the network, how it is changing and networking technologies – not social networking or web marketing firms like Groupon; my prior GRPN comments are in this post.

At time I was writing my thoughts on how the network is changing, what I was not thinking about was how Arista Networks is affecting the ecosystem inSilicon Valley.  If you are looking for an analysis of Arista Networks, this is not it.  I am thinking about Arista, but this is not intended as a technology or solution overview; maybe I will do that for another day.

To understand what I am thinking about, we need to go back to Nuova Systems.  Cisco acquired the company in April 2008.  From this company emerged the Nexus 5000 class set of products.  What was unique about the company was the funding and the impact it has had on others in the Cisco ecosystem.  Think about where Cisco stood in 2008.  For the FY2008, Cisco sold $13.46B in switching products.  (Note that Cisco’s fiscal year is August-July).  FY2008 was the peak year in Cisco switching revenue before the onset of the global credit crisis in September 2008.  I have detailed the 2009 catch-up spend event here and here, but consider the world before the credit crisis induced spending collapse and subsequent catch-up phase.  Cisco was the dominate market share owner in switching.  There was no one close and they had spent their acquisition capital from the start on this market segment.  This is a just a partial list of Cisco’s switching centric deals starting the most important being Crescendo in 1992.  This list alone totals $6.2B:

  • 24-Sep-93 Crescendo Communications     LAN switching                    $94,500,000
  • 24-Oct-94 Kalpana                                              LAN switching                    $204,000,000
  • 6-Nov-95 Grand Junction Networks           LAN switching                    $220,000,000
  • 6-Aug-96 Nashoba Networks                         LAN switching                    $100,000,000
  • 3-Sep-96 Granite Systems                               LAN switching                    $220,000,000
  • 31-Aug-99 IBM Networking HW Div          Computer networking      $2,000,000,000
  • 11-Apr-00 PentaCom                                       LAN switching                     $118,000,000
  • 20-Aug-02 Andiamo Systems                      Datacenter Switching       $2,500,000,000
  • 29-Jun-04 Actona Technologies                 Data storage                         $82,000,000
  • 8-Apr-08 Nuova Systems, Inc.                    Datacenter Switching        $678,000,000

Total  $6,216,500,000

Nuova was started in mid-2005.  It was self funded by the Cisco executives who left to form the company. This list included: Mario Mazzola, Luca Cafiero, Prem Jain, Soni Jiandani as well as outsiders Ed Bugnion and Tom Lyon.  In August 2006, Cisco increased their ownership to 80%.  In April 2007, Nuova had ~200 employees and Cisco agreed to a cap of the potential buyout of remaining 20% at $678M.  A year later, April 2008 Cisco closed the deal to acquire the remaining 20% for $678M.  No venture capital firms were involved to my knowledge. Jayshree Ullalleft Cisco in May 2008.

Back to the list of Cisco acquisitions from the 1990s; Granite Systems was founded by Andy Bechtolsheim and David Cheriton.  In 2008, Andy Bechtolsheim became Chairman and CDO at Arista networks, which he had founded a few years earlier with David Cheriton.  Jayshree became CEO in October 2008.

In October 2008, Sequoia Capital produced a presentation for their portfolio companies entitled “RIP Good Times.”  It should be noted that Sequoia was the first and I think only VC to invest in Cisco before the IPO in 1990.  In the month of October 2008 these events are occurring:

1. Cisco is having their best switching quarter in their history with revenues closing at $3.54B ending the fiscal year at $10.4B.

2. Sequoia (original Cisco VC) is telling their portfolio companies to cut burn rates, conserve cash, reduce expectations and go into the bunker.

3. Arista namesJayshree UllalCEO. They are just a few months away from their first sale, focused on low latency, high performance compute clusters against Infiniband.

4.  Juniper is a few months into the development phase for QFabric.

Did I mention that like Nuova, Arista does not have any venture capital firms as investors?  At least to my knowledge there are no VC investors.  My prediction is we will look back on October 2008 as a pivotal month in the history of networking.  The few venture capitalists focused on networking dug in for a long winter’s nap.  Their thesis was why be invested in a market in which Cisco is the dominant company?  It is also a market that requires a high level of capital to be successful and is perceived to be a mature market – all of which describes a poor set of conditions for entrepreneurial companies.  It is much easier to invest in mobile apps and web properties, than fight incumbents that have scale and capital advantages.

A small group found that the darkest of times was not the time to sleep.  It was the time to plot a revolution and start a new company.  While the powerful were distracted – others saw opportunity.  That is what was happening in October 2008.  As I have written in the past, we are on the verge of significant shift in the structure of the network.  It is not clear who the winners will be.  The manner in which Arista is affecting the venture capital community is in the terms of lost opportunity.  Arista is aggressively marketing the company to the investment community and I suspect they will have large and successful IPO in 2012 – an IPO in which no VCs will participate and no LPs will receive shares.  I have seen only one other early stage startup in the datacenter space.  There are two others that I know of marketing an A round, but the game is afoot and the clock is ticking.  This is not some Openflow appliance game; I am referring to a big league game for real estate in the datacenter – the skyscrapers of the future.  If this was 1992 or 1998 there would be 6-10 players in the space – but it is 2011 and it is hard to find players for the big game.

I wrote this in early 2006 and without being around any of the events in October 2008, I would hypothesize that this is what was happening.  “The victors in any revolution are the companies that are led by great leaders who understand the cycle of change and strategize not for victory on the day of revolution, but for victory in [end].  When Microsoft becomes an old regime and the conditions required to support a new revolution materialize, revolutionaries will come forth.  Even today there are revolutionaries who operate in fear of Microsoft.  Similar to the formative years that shaped a young Bill Gates; the new revolutionaries might be students at Harvard, MIT, and Stanford or located in far away places such as Moscow, India, and China.  They might even be employees of Microsoft.  At night they clandestinely plot the downfall of Microsoft in darkened basements of their homes, fearing that Microsoft will not discover their treacherous plots for they know it will cost them their jobs.  This is where revolutions begin.  They begin with the revolutionary who is motivated by the quest for change, the desire for power, and the belief in a higher ideal, a better plane of existence that in the end will provide some form of social justice and personal satisfaction.  Revolutions start when people create a solution to a complex problem.  The people working inside Google are clearly revolutionaries who are planning to attack Microsoft from outside the computer operating system.  Google’s strategy looks very much like the new generation of service providers who want to offer a service without owning a network.  Google is looking to create applications that reside in the web and do not require a specific operating system.  If Google can find a way to bypass Microsoft’s core advantage in their dominance of the personal computer operating system and productivity application markets, it will mean that Microsoft will have a major revolution to endure.  The United States is born from revolution.  We have adopted an economic structure that is not perfect, but it fosters change.  The United States changes its government every two to four years.  We have a predicable timeline of points when change might occur, called elections.  We have an economic system that rewards success and punishes failure.  We have a private equity mechanism in place that encourages risk, supports revolutions, and rewards success.  “…in America you can drop out of college and start Microsoft, Oracle or Dell.  You can get a “C” grade at the Yale School of Management and still launch FedEx.  You can dropkick your Ph.D. pursuit and start Google,” [see Rich Karlgaard, Why we Need Startups, Forbes, July 21, 2003].  The result is a nation-state that is far from perfect, but has an infrastructure in place that supports change and revolutions.  The roots of this system can be traced back to the Sons of Liberty in Boston during the years leading up to the American Revolution.  The people that confronted the government in London found their financial backing from private funds within the American colonies.  Privately raised funds were not being used to start a new biotech startup on the banks of the Charles River in 1775, but they were being used to fund the active opposition to General Gage, his military forces in Boston, and the English Monarchy of King George the III.  This is how wealth, revolution, and ideas are linked.”


* It is all about the network stupid, because it is all about compute. *

** Comments are always welcome in the comments section or in private.  Just hover over the Gravatar image and click for email. **


End of the Era of Large Scale Venture Backed Innovation?

The years 1964 and 1965 mark important milestones in the history of computing and telephony.  In 1964 AT&T opened the first central office in the United States.  This was the beginning of the migration from operators and five digit dialing, to a nation-wide deployment of circuit switching technology and ten digit dialing.  In 1965, IBM introduced the S/360 mainframe computer.  When I think of the big bang moment for the telecom, computer and internet industries I think it started in 1964-65.  We have gone through 45 years of distributing the power of the mainframe and the CO into the network and users.  Now we are at a very interesting point in the network and I think it would foolish to assume that the distribution of computing and networking power will recede.  I think it will accelerate.  The hypothesis I am exploring is: we are at the inflection point in the network that was dreamed about and theorized about ten, twenty years ago and now that we have arrived, is the market only accessible to public companies capable of large scale R&D efforts?

I have spoken to many a VC who thinks that innovation rarely occurs in large companies, yet I would offer up as counter argument that it is true that large companies become trapped by their legacy products, but it was large companies that gave us the MAC, Razr, iPod, iPhone, iPad, DQPSK, Pentium, InkJet…we could write a very long list.

Some Data to Consider

I looked at some of the data from the National Venture Capital Association as well as a model I created of annual revenues for 18 companies.  I chose these companies at random and maybe if I was sitting at a Factset terminal I could produce a broader, more complete model, but I was just looking for a mix of large cap to small cap, networking, storage, telecom and enterprise focused companies.  The companies in the model are:  CSCO, JNPR, Alcatel, LU, NT, BAY, CIEN, BRCD, ERIC, NTAP, INFN, APKT, EMC, ARUN, TLAB, FFIV, RVBD and ADTN.  I left out HPQ, DELL, AAPL and many more companies.  My selection was random as this is a blog – not a market study.

Chart 1 below shows the combined revenue of these companies from 1995-2010.  It also shows the investment in the networking and equipment sector on an annual basis from the NCVA.  I also marked what I termed as the Lost Decade for VCs from an earlier post.  If you did not read that post I describe the period from 2001 to 2010 as a lost decade because many of business plans from 2001 are now valid, one simply needs to erase a zero and make 2001 read 2011.

Here is chart 2 in which I added the VC investment in internet specific companies.  I understand that the capital intensity of internet investing is far less than hardware companies or real science based investments.  I would note that the creation of internet specific companies are drivers of network usage and thus can be considered a growth engine for network infrastructure they are using and the creation of long form content.

Here is the third chart and I extrapolated some expected growth rates for the 18 sample companies into 2015.  Note that even the global credit crisis in 2008-2009 is a blip.  As VCs have been avoiding the networking and equipment market, revenues have been going up and the drivers of the new network are going to cause an explosion in network usage and investment.

Concluding Thoughts

When I look at these charts I talk to entrepreneurs and VCs I am increasingly convinced that there still is a hangover effect from the 2001-2002 investment cycles, but I also see increased complexity and cost in the product cycle for system level products.  Can venture backed companies still innovate in system level networking or is it all about low cost web based models that take $10M of capital to provide a proof before deciding to scale out the business?  Is large scale venture backed innovation dead?  If you believe in the thesis that Carlota Perez created about markets, then we should be in a new golden era for networks.  Here is reproduction of her model:

In 2006 I produced the following chart as part of manuscript I wrote thinking about Perez’s theories.  I dug it out and updated the model below.  It is reasonably accurate, but a few corrections are in order.  I called what became known as cloud computing, grid computing.  I had the content model correct, just not the all players.

At a recent lunch meeting I had an interesting thought expanded upon by a person who has been in a number of VC backed startups.  We were talking about the how the process of creating a horizontal company along the lines of the World is Flat theory enabled companies like Huawei to get where there are today and that there is an accelerating trend towards vertical integration (back to the past) in technology.  More in house silicon than you think: JNPR, AAPL, CSCO…the list goes on.  Companies want to protect IP and move away from homogenized designs using common components, merchant silicon and they way to do that is either through software or custom silicon.  For the networking companies that is where the future lies and even Huawei knows it.

I think the opportunities in the new network are massive, but I also think many investors are going to miss it because the capital required to be in the market is still handcuffed from 2000-2001 collapse and the lost decade.  The drivers are omnipresent: 100G, LTE, Real time content, Long form (or big data) content, Cloud Computing, Persistent connections and all photonic switched DCs.  It appears that VCs are going to leave these markets to public companies.


* It is all about the network stupid, because it is all about compute. *

** Comments are always welcome publically or in private.  Just hover over the Gravatar image and click for email. **