It’s been a fun few weeks in the world of IP routers and networking equipment, as the two of the most dominant players — Cisco and Juniper — surfed the chaos that is known as the telecommunications market.
Let’s explore what this means, and particular what it means for the IP networking market. But first, let’s cover the soap opera.
First up: Juniper CEO Shaygan Kheradpir was abruptly dismissed from a nicely paid corner-office job, for conduct unbecoming a Juniperian. Juniper’s press release was minimalist, hinting only at “his conduct in connection with a particular negotiation with a customer.” My sources indicate this was related to way he handled the Verizon account, which, ironically, is where he used to be employed as the CTO. Tongues are still wagging in Sunnyvale, Calif., as insiders and outsiders speculate about what specific action would drive the the exit of a major-company CEO in under 12 months after taking the job.
Next up: AT&T, one of the largest service providers in the world, slashed its capital spending plans by 15 percent in concert with buying Mexican mobile provider Iusacell. Coincidence? I don’t think so. Money in, money out. AT&T’s got to pay for its high priced deals some how, and clearly the pesos came straight out of the capital budget.
Next up: Cisco’s closely watched fortunes showed some schizophrenia again, as the company reported quarterly earnings that beat expectations, yet it guided down on future revenue. Shares got a bit of a boost since dark news was already baked in, especially since AT&T had announced its capital spending downgrade just a week earlier. But what hints did this give us that Cisco is back on the right track? Not many, really. The forward guidance was uninspiring, and there’s no evidence that the slowdown in the emerging markets and service provider markets are going to boost Cisco anytime soon.
As Wall Street analysts, business media, and activist investors play point-counterpoint about the future prospects of Cisco and Juniper, I want to ask deeper, philosophical questions, such as: 1) Do you really want to sell to service providers? and 2) What’s the future of IP routers?
Data Center Tidal Wave
Both of these markets, IP routers and service providers, are emblematic of a tidal wave of change and shifting dynamics brought about by cloud computing. Cloud has shifted networking architectures away from enterprise networks and service providers and into the cloud middlemen — a vast collection of data centers powering web apps across the globe. The bulk of the money going into data centers is being invested in Ethernet switching, server gear, and virtualization software.
It’s also well-publicized that some of the most cutting edge data centers, such as those at Google and Facebook, are relegating Ethernet technology to commoditization and even combining the functionality with servers that can do all the networking with software, in a trend we paint with the broad brush of software-defined networking (SDN).
What that all boils down to is really tough markets for any company selling networking products either to enterprises or service providers — and a fierce demand to battle for the heart and soul of the data center. There’s a reason why VMware paid more than $1 billion for Nicira and Cisco fired back with nearly $1 billion for Insieme. They both know what data-center virtualization is where the money’s at.
The drama at Juniper, and to a lesser extent, Cisco, reflects this anxiety. The world has changed, and these executives are desperate to cling to to their older markets while the uncertainty of the new markets unfolds. I would imagine that the seas are about to get rougher, as service provider capital spending gets whittled down and the rise of the SDN data-center server/switch begins to accelerate. Cisco and Juniper, long-time kings of networking, are losing all of their traditional sources of strength.
Service Provider Stagnation
So, I believe that customer trends indicate that Ethernet is becoming a more popular, less expensive solution than routers for many applications. However, the market is far for straightforward. If you looked across different segments and geographies, according to many large market research services, it’s more complicated.
For example, IDC points out that Ethernet switching was growing the fastest in Asia/Pac, growing 21 percent year-over-year in the second quarter, and overall it grew 6 percent. The worldwide router market declined 4 percent year over year in the second quarter of 2014, with varied results across the key segments within the market. The service provider segment declined 0.9 percent year over year while the enterprise segment was down 12.4 percent year over year.
Dig into the numbers further and you find another trend: Both routing and switching growth is flat or declining in enterprise markets, but growing in data center markets. And that’s where Ethernet growth is even stronger.
All of these swirling trends, though complicated, tell you a lot about the market dynamics of networking and the kind of trouble that Cisco and Juniper are up against: The service provider market is stagnant or declining, Ethernet switching is in many cases replacing routers, and most of the action is in the data center. This may be why Ethernet switches hold an advantage over IP routing as a market to watch: The cloud computing boom has shifted the architecture of networks to create more large-scale switching centers, diminishing the need for enterprise routing.