Switch Wars (At An Interop Near You)

This week the networking industry will attend the annual Interop trade show in Las Vegas.   While Software Defined Networking (SDN) and its potential disruption or new growth driver to the networking industry continues to be the main hot topic in the industry, the Data Center Ethernet Switch market remains a vibrant, good growth market in the near/intermediate future.  SDN will clearly be a major force in the networking industry over the next several years, but winning Ethernet Switching Data Center customers and footprint today will only enhance a networking company’s chance of being relevant and driving the transition to SDN.

In the past week, Data Center switch vendors Arista, Brocade and HP all announced new Data Center Ethernet Switch products.  In addition, Juniper announced its new Data Center switch about a month ago.  These and other competitors such as Extreme, Dell, and Huawei are all aiming to be of significant size and the number two Data Center Switch vendor after Cisco, which clearly dominates this market with over 65% market share.  In looking at all the recent Data Center switching product announcements, it seems to me that Arista with its new 7500E continues to stay ahead of the Cisco challengers when looking at a variety of key metrics such as latency, scalability (i.e. supports up to 100,000 servers in a two-tier design) and interface port flexibility (i.e. single line card for 10/40/100GbE).  While these hardware metrics suggest Arista continues to stay ahead of its peers, Arista also was early in developing an open based operating system based on Linux, which positioned the company’s Data Center switch products on the SDN path well before SDN was a hot topic in the networking industry.   In simplest terms, Arista seems to be given the networking industry what it wants, namely, a strong technical product line that offers both best in class hardware metrics and an open and easy to use operating system.

While the battle for number two after Cisco in the Data Center switch market is far from over, especially given the larger resources some of Arista’s competitors possess in terms of R&D, distribution and product breadth beyond switching, Arista seems to me as the one to beat.  I also think Arista’s technology product momentum is making them the likely company to beat in the high profile NYSE Data Center Switching “jump ball” that has been ongoing on for several months.   From a stock perspective, if Arista (currently a private company) can continue to leverage its technical advantages in the Data Center Switch Market through increasing market share, its success will likely take away a key potential stock catalyst for its publicly traded competitors.   While Cisco will always be the number one in Data Center Switching, a strong, rapidly growing number two could make for a good investment opportunity.  Right now, Arista is making that challenging for its publicly traded peers.

Disclosure: I do not currently own shares of any company mentioned in this blog post.  NT Advisors LLC may solicit any company mentioned in this blog post for consulting services.

What (if Any) Part of the Networking Value Chain Will Be Disrupted by SDN?

I have been following the topic of Software Defined Networking (SDN) for the past three years.  Three year ago the technology was not well known by Wall Street but now is enjoying an intense level of discussion by investors. When the technology was first presented to me three years ago, my initial reaction was SDN would be a risk for technology companies in the Ethernet Switching and Routing markets (e.g. Cisco and Juniper), while creating new opportunities for semiconductor companies selling merchant silicon (e.g. Broadcom and Intel) and newly created SDN software companies.   After visiting a few SDN private companies in the past couple weeks, talking to industry participants and reviewing recent SDN acquisitions by Cisco, Juniper and others, it is actually less clear to me now how SDN will dislocate the current networking value chain.  I am not questioning the value proposition of deploying SDN or the likelihood that it will be a significant investment cycle in the next five years, as I view that as a given.  Rather, the question whose answer has become less obvious to me is which publicly traded companies (if any) are most vulnerable to the upcoming SDN technology cycle and when will this dislocation most likely begin being reflected in these company stock prices.

– Will the ultimate acceptance and deployment of SDN match my simple initial reaction that it will be negative for Cisco and Juniper as switching and routing face competition from more open oriented hardware platforms (Arista and Pica8 are examples of privately held open hardware platforms)?

– Will SDN actually require more complex and high performance hardware platforms in the data center as the real value around SDN will be operational simplicity and cost reduction rather than a focus on hardware costs (data center privately held platform companies include Arista and Plexxi)?

– Will SDN be more of a risk for Layer 4-7 companies that are selling special purpose appliances that may be made obsolete by more multi-functional and integrated software applications in the SDN orchestration layer?

– Will both Layer 2/3 and Layer 4-7 companies be at risk?

– Will SDN create the opportunity for a Network As A Services (NaaS) model and disrupt the entire networking value chain?

– Perhaps, SDN will be some combination or elements of all the above scenarios?

I will be moderating an investment panel at an SDN user conference in a couple weeks with some very smart and experienced investment professionals and hope to get more insight into these topics, which I plan to share on this blog.  In the meantime, lets take a look at how sentiment on SDN’s impact to current publicly traded companies has changed over the past several months and why it is likely SDN developments will not be that relevant to public company stock prices in 2013 as they were in 2012.

The first major wake-up call on SDN to the public markets was July 23rd, 2012, when VMware announced it was acquiring Nicira for $1.26 billion.  Since this announcement occurred after the market close, I was curious to see how Cisco and VMware would trade the following day.  As it turned out, Cisco’s stock lost 6% of its value (about $5 billion in market value) the next day while VMware fell about 0.3% (note VMware announced earnings the same evening it announced the Nicira acquisition which likely muted the impact of the Nicira acquisition to VMware’s stock price movement the next day).  What was interesting about the market reaction was that Cisco lost $5 billion in value while VMware barely budged after spending $1.26 billion for a company that at the time was likely to generate less than $50 million in revenue in calendar 2013.   Clearly, the market at that point viewed SDN as a massive technology risk to Cisco.

Over the course of the next several months, however, Cisco formulated its SDN strategy, made a couple of SDN acquisitions of its own (vCider, Cariden and funded Insieme) and communicated its SDN strategy at its analyst day on December 7th, 2012.  Juniper acquired SDN start-up Contrail and communicated its SDN strategy on January 15th, 2013.  In addition, Nicira/VMware seemed quiet in terms of market penetration and customer deployments post the announcement of the acquisition in July.  So, in the span of 6 months, SDN went from a perceived significant risk factor to Cisco and Juniper to being more of an unknown entity both in terms of potential impact and timing of that impact.  Investors slowly began to realize that SDN would have little impact to 2013 and maybe even 2014 financial results.  Also, Cisco and Juniper are fighting back and will aggressively try to leverage their installed base of equipment to take advantage of SDN as a new revenue opportunity.

Now lets look at Layer 4-7 (e.g. security, load balancing, application delivery control).  What I find interesting here is several of the new SDN private company fund raising in the past several months were for companies attacking this segment of the networking value chain.  Companies that might fall into this category include Embrane, LineRate, PLUMgrid and Pluribis.  Several industry people I speak to suggest that Layer 4-7 will actually be the first area of SDN deployment in data centers given the need to provision and manage policies/applications/security at scale in the data center, which proves to be difficult when managing multiple single purpose appliances and that managing this in the orchestration layer within the SDN model potentially provides an elegant and scalable solution.   It might be coincidental, but in listening to the F5 earnings call this week, I found the following dialogue in the Q&A portion of the call on why F5’s Technology Vertical has not been performing well in the past couple of quarters very interesting as it relates to this topic. Below is how F5 management responded to this question:

“So, on the Tech Vertical issue, you’re right. I mean, the Tech Vertical has trended down over the past several quarters for us, and we believe it’s driven really by a couple of our larger customers that are taking alternative architectural approaches in terms of how they’re building things. So, generally they’re building some basic functionality into that app. And, so we’ve been seeing that going on, and obviously we’re doing something about it.

We’ve got projects going on internally that we believe will provide this type of customer with ways that will make it easier for them to integrate our functionality into the applications [inaudible] that they’ve got.”

Source: Seekingalpha.com

What is interesting here is that the Technology Vertical within F5 results typically includes major data center and cloud providers in the category of Facebook, Apple, Google, Yahoo, etc.  While I do not know which specific customers F5 was referring to in this comment, it is valuable to see how such large-scale operators are already implementing certain parts of the Layer 4-7 stack on their own.  One can easily infer why Layer 4-7 SDN start-ups are getting funded at a nice clip given the potential for disruption here. Obviously, publically traded Layer 4-7 companies are not standing still as this is happening and are already offering virtual instances of their appliances, which I would imagine will ultimately be offered as applications in the SDN orchestration layer.

Finally, start-up Pertino appears to be focused on using SDN as a framework for Network As A Service (NaaS).  While they are not likely to be the only company pursuing such a business plan (perhaps some of the companies mentioned above), it does the raise the option that NaaS could be disruptive to the entire networking value chain especially if we see large players like Amazon, Google and others pursue such an offering or if a new disruptive start-up emerges to be the Saleforce.com of NaaS.

So in summary, SDN it is going to be a very disruptive technology. What was initially viewed as a technology shift that will be a negative for Cisco and Juniper is now potentially more complex to predict in terms of public market investing.  What is likely, however, is that SDN will have little impact to publicly traded stocks in 2013 as other macro and company specific fundamentals will be more relevant to stock prices in my view.  I doubt we will see another VMware/Nicira type of deal in 2013 both in size and in its impact to publically traded stocks like the $5 billion in lost value Cisco experienced the day after this deal was announced.  However, over the course of the next year or two, the potential impact of SDN to publically traded companies and how these companies either capitalize or fall victim to the adoption of SDN will be more evident.  It will certainly be fascinating to watch!

Disclosure:  I currently own shares of Cisco mentioned in this blog post.  NT Advisors LLC may currently or in the future solicit or have as clients any company mentioned in this report.

Juniper Stock Seems Stuck

After reviewing Juniper’s 3Q earnings release and listening to their earnings call tonight, I came away continuing to think that Juniper’s stock is stuck in many ways.  In particular, the company seems to be bound to single digit revenue growth at slightly over $1 billion in quarterly revenues and a peak quarterly run rate of $0.20-$0.25 in pro-forma earnings per share.   In the world of technology stocks, former darling growth companies that seem stuck in terms of revenue growth and peaking earnings power suffer the ongoing melting away of their valuation multiple. Clearly Juniper has seen their valuation compress over the past several quarters, but hopes of an expanding multiple do not seem likely for now.   With earnings bound  between $0.80 and $1.00 on an annualized run rate, the stock is not likely to get over $20/share in my view anytime soon.

The key for Juniper’s valuation expansion will be acceleration of its revenue growth and margin preservation. While new products like the PTX in MPLS Core Switching and Q-Fabric in Data Center Fabrics have provided hope to the Juniper bulls on the stock over the past year or so, these products continue to ramp very slowly.  There also continues to be doubt on how successful the Q-Fabric product will be given its lack of significant revenue traction in the past few quarters, increasing competition in this segment of the market and the potential overhang that the open standards Software Defined Networking (SDN) architecture poses to the originally closed Q-Fabric offering.  Juniper mentioned on their call that SDN is a key area of focus for the company, and it is likely the company will focus its efforts in the future to make the Q-Fabric more open and less reliant on the originally planned closed Juniper Q-Fabric controller.

There also continues to be a lack of excitement in the future growth prospects of the traditional Router and Security businesses.  Both Routing and Security products have shown year over year declines for the nine-month period through the end of 3Q12.  One could argue somewhat convincingly that routers are suffering a cyclical decline and are poised to recover at some point, but even so, the future growth is likely to remain sub 10% in the future.  Security is not suffering any cyclical decline, but rather a case of niche specialized competitors gaining share (e.g. Palo Alto Networks, Fortinet etc…) as Juniper has lost competitive advantage in this market.  \ It is unlikely this business will return to a consistent growth story in the future.  The company continues to do well in the traditional enterprise Ethernet Switch market growing 20% year to date vs. last year, although this business tends to be lower in margin profile than both routing and security, thus, not likely to help the long term business model.   Even so, this is the one share gainer Juniper has going for it right now.

What Should Juniper Do?

Juniper is in a multi-front war in Service Provider Routing, Enterprise/Data Center Switching and Next Generation Security Firewalls all of which will also be challenged in some unknown way by the emergence and deployment of SDN in the next 3 years.  The status quo of investing in all of the above seems like a high risk proposition for Juniper as it will have to continue to fend off much larger companies like Cisco and Huawei and smaller product specialists like Arista Networks, Palo Alto Networks etc…  My sense is Juniper will need to focus more, get back to its roots in the Service Provider market and seek larger committed partners that can help it succeed in the Enterprise market.  Breaking up the company while likely very difficult could also be a favorable outcome for shareholders as different buyers would emerge for the Service Providing Router business and the Enterprise business.