This morning SDN Journal broke a story stating that Dell is off to create its own SDN consortium (separate from OpenDaylight) with potential partners like VMware, NTT, Google, Microsoft, and Big Switch. While we don’t have first hand knowledge to know if this is true or not – it really doesn’t matter – because the idea of Dell (or others in this cast of characters) attempting to build a second alternative doesn’t make customer or business sense. Here’s why:
- Dell’s market position. Dell has small market share in networking – how do they have the clout to drive customer adoption of yet another open-source SDN controller. They don’t. No other large networking companies are listed as potential partners of this consortium – so does that mean the only network hardware that will support a Dell backed controller is Dell’s? How do two competing projects (Dell’s and OpenDaylight) align on a standard, such as, OpenFlow while building separate code bases? Feels hard. What incentive will participants of Daylight have to work with yet another controller? Very little.
- Wrong cast of characters: What is the economic incentive for Google and NTT to drive an initiative for another SDN controller? It will be easier and cheaper for them to just fork Floodlight or OpenDaylight and write their own. The second data point that doesn’t fit is Microsoft, Google, and VMware banding together from a cloud services perspective; they are bitter enemies and I don’t yet see economic or market activities that would drive these three to band together in networking – at least today.
- Competing Controllers: This is the same argument and the same knock some vendors throw on OpenDaylight around the politics of what controller is used. If the Dell, etc rumor is true, which controller is the code base? Our understanding is that Google, Nicira (VMware), and NTT all own rights to the Onyx code base that supposedly forms the basis for both Google’s home grown controller and Nicira’s controller. How does that get reconciled with Big Switch’s two controllers (Floodlight and Big Network Controller)? Or NTT’s own controller technology? Sounds just as messy if not messier than Daylight
Looking at this from another side of the coin: given that, last we heard, Michael Dell is a personal investor in Big Switch and concurrently attempting to remake Dell into a major networking player, I would be surprised if Dell and others haven’t considered (i.e., held a meeting about) if an alternative to OpenDaylight was feasible. Though from our perch – we can’t see how this collection of companies could create something that is viable for both customers and developers.
To state our position — we are neither pro or con OpenDaylight, Floodlight, or other open-source controllers or consortiums. We are pro customer — what pro customer, is a viable open-source alternative to proprietary solutions. The question is will any of these open-source approaches provide a viable alternative (where viable is defined by mainstream customers adopting it for their business). It’s too early to tell on Daylight — and like we stated before, we don’t believe the current path for Floodlight is viable from an open source perspective.
Who knows – maybe my logic is flawed and we’ll see a new open-source controller on Tuesday – or we won’t. Ether way, we’ll learn if it’s me or this rumor that is silly….should be fun.
Check out more OpenDaylight on SDNCentral:
- Juniper Warms Up to OpenDaylight
- SDNCentral SDN and NFV Weekly Roundup — April 11, 2014
- SDNCentral SDN and NFV Weekly Roundup — April 4, 2014
- Q&A + Video: Plexxi Demos Application Abstractions with OpenDaylight and Real Status Visualization
- Network Operators Love Open-Source and Open Systems — in Theory