Cumulus Networks (B): Competing with the Channel
In 2013, months after coming out of stealth, Cumulus Networks had a problem many start-ups would envy: hardware makers wanted to sell its product. The Mountain View, California, company sold Cumulus Linux, software that let companies run data center networks on inexpensive switches from almost any manufacturer, an Android-style alternative to the integrated, iPhone-style approach of incumbents such as Cisco. But open networking required integration work that most corporate IT departments could not do on their own. Partners could close that gap, at a price: a share of every sale, plus months of training, joint selling, and engineering support from a small team. Three suitors stood out. Dell offered the strongest brand and broadest enterprise reach but sold networking software of its own. Edgecore, a Taiwanese maker of low-cost bare-metal switches, would do almost anything Cumulus asked but could generate little demand. Mellanox had real sales reach and was pushing hardest, but supporting its chips would cost millions and take more than a year. Cofounder and CEO J.R. Rivers had to decide whether to rely on partners at all, which ones to back, and how to tell whether a partnership was working.
The (B) case picks up years later. Customers still found open networking harder to buy and support than an integrated system from an incumbent, and stacked vendor markups eroded much of its price advantage. CEO Josh Leslie and Kaushik Shirhatti, vice president of global sales, decide to launch Cumulus Express, a Cumulus-branded switch sold as one integrated product, but have the capacity to build it with only one hardware partner. Whichever partner they choose, Cumulus will be competing with its own channel, and the choice will send a message to the company Leslie sees as Cumulus’s most likely acquirer. A short (C) epilogue describes what happened.