Cisco Systems’ impressive earnings performance this week begs a question: why has Juniper Systems annual revenue somewhat stalled? For the past few years, Juniper has seemingly been playing catch-up relative to the heavyweight enterprise networking solution providers. A history of non-accretive acquisitions, multiple runs at building out a Wi-Fi portfolio, misses in delivering software-defined networking (SDN) capabilities, and difficulty in driving a sustained open source/API-led development plan are all contributors in my mind. I’d like to dig a little deeper into these areas.
Some acquisitions that simply didn’t pan out
In 2004, Juniper acquired NetScreen Technologies for a staggering $4B in stock, in an attempt to shore up its security portfolio. Granted, a lot has changed in the last 15 years with the vulnerabilities created by IoT sensors and the steady escalation of threats from both outside and within network environments. However, customer confidence in Juniper was shaken a few years ago with the discovery of unauthorized code dating back to 2012 that served as a “backdoor” for decrypting traffic (documented by CBS News here).
In 2010, Trapeze Networks promised to bring controller virtualization to wireless networks at a price of $152M, but Juniper abandoned the platform for partnerships with Aruba and Ruckus Networks (eventually announcing its intent to acquire Mist Systems this past March). In 2012, the company acquired Contrail Systems for $176M in an effort to boost its SDN capabilities. Five years later, though, Juniper handed what became OpenContrail off to the Linux Foundation when it couldn’t build momentum behind a developer community.
Is open source a priority?
In recent press from the Open Infrastructure Summit that wrapped up earlier this month, Juniper spoke of its participation in the open source community. After its transfer of OpenContrail to the Linux Foundation, the project was renamed Tungsten Fabric. Juniper still remains involved today. An executive stated in a recent SDxCentral article that can be found here that the various open source projects still face challenges with customer adoption. Maybe that’s based on Juniper’s historic focus on the carrier/telco and service provider side of the market vs. enterprise. However, in the time I’ve spent at Linux Foundation user group events such as the DPDK Userspace and Open Network Summits in North America and Europe, I’ve witnessed firsthand adoption across both domains (if interested, you can find my most recent article on that topic here).
It’s not my intention to pick on Juniper. I think it has told a good story but has fallen somewhat short on the execution. I know the executives at Mist Systems, and they do bring some interesting intellectual property to the table for Juniper. I was invited to participate in a phone briefing with Juniper and Mist shortly after the acquisition announcement, and I applaud executives in allowing Mist operational autonomy to continue its innovation efforts around AI and Wi-Fi. However, I would like to see Juniper invest more resources in building out a stronger SDN offering as well as a developer community with an API-centric approach. The latter will not only benefit the company’s efforts in open source but also foster innovation on top of its hardware. One only has to look at the success of Cisco’s DevNet and HPE Aruba’s Airhead communities for examples. These infrastructure providers are creating value-add and “stickiness” with channel partners; perhaps Juniper could take a page from their playbooks.