Remember CloudNFV? The 2013 NFV group was ahead of its time -- maybe too far ahead. It was an ambitious initiative to run virtual network functions in the cloud that got off to a promising beginning, got handed off to Dell leadership, and then withered away.
CloudNFV was an industry group founded in 2013. While NFV has made huge strides since then, CloudNFV's mandates look enormously ambitious, even today. As the name implies, CloudNFV's mission was to focus immediately on running NFV in the cloud, skipping the intermediate step of running NFV software virtualized on servers.
Leading advocates today espouse the same vision for "cloudifying" NFV that CloudNFV advocated two years ago. But today's advocates see cloudification as an end-state, achieved through hard work by leading-edge comms companies partnering with vendors. CloudNFV wanted to jump through the beginning stages and go straight to the end. And they had that vision two years ago -- not that long ago by real-world standards, but ancient history in the fast-moving world of NFV.
Headed up by Tom Nolle, veteran industry consultant and president and CEO of CIMI Corp., CloudNFV launched in July 2013, with six vendors in the group. The first project was a cloud-based implementation of the IP Multimedia Subsystem (IMS) for supporting converged services and enabling network functions to be broke into reusable modular segments. (See New Group Ties NFV to the Cloud.)
By November of that year, Nolle told my colleague Carol Wilson that the CloudNFV software would be productized and appearing in network operator labs by 2014. The group demonstrated its cloud-based IMS implementation in October. (See CloudNFV Moves Quickly to Product Stage.)
The ETSI/NFV ISG approved CloudNFV's work as its first proof-of-concept application in December 2013, with service provider sponsors Telefónica SA (NYSE: TEF) and Sprint Corp. (NYSE: S) Vendor sponsors were 6WIND , Dell Technologies (Nasdaq: DELL), EnterpriseWeb LLC , Mellanox Technologies Ltd. (Nasdaq: MLNX), Metaswitch Networks , Overture Networks Inc. , Qosmos and Aeroflex Inc. (Nasdaq: ARXX). It completed last year. (See NFV Group Kickstarts Proof-of-Concept Work.)
A few months later, Dell had taken over leadership of CloudNFV, with that company's Wenjing Chu, distinguished engineer and CTO in mobility and net virtualization, replacing Nolle. Chu told Light Reading that CloudNFV was part of a "pan-Dell" approach to virtualization that would span several years, with a "big announcement coming out very soon about a corporate-wide strategy." (See Dell Has Big NFV Plans.)
But since then, we haven't heard anything about CloudNFV.
"CloudNFV never really ignited after I left the project," Nolle tells Light Reading this week. "Part of that, I think, was due to Dell's desire to focus on selling infrastructure rather than on building a total solution. I retained the trademark and website because I couldn't be sure that Dell would continue the project as it was, and while others registered similar domains there was never any activity. I promised to link to any project website if a URL was provided to me, but none was."
Nolle says he worked on CloudNFV for free at first, but there was no interest from Dell and other companies "to pick up and pay when the free stuff ran out."
From Dell's perspective, CloudNFV ran its course. As Nolle says, Dell is focused on selling infrastructure -- servers, storage and networking -- to comms companies, letting partners, including VMware Inc. (NYSE: VMW), Overture and others, sell the NFV software itself. CloudNFV work taught Dell how to tune its equipment to work with NFV software from its partners, Jeff Baher, senior director, NFV strategy and marketing, tells Light Reading this week. "Anyone that's going to provide a common platform needs to understand the hardware and software and anything that runs on top of that," Baher says.
Sources in Silicon Valley tell us CloudNFV was part of a Dell experiment to develop its own NFV software stack. But Dell decided not to go that route, focusing instead on selling servers, storage and networking, and working with other companies on the NFV software.
And that's where the lost opportunity comes in. Or possible lost opportunity. Dell could have transformed itself from an infrastructure provider to selling full solutions. But it decided not to go that route, focusing instead on its core infrastructure competency rather than risk alienating partners by competing with them on software. The recent announced intent to acquire EMC Corp. (NYSE: EMC) is an extension of that strategy -- EMC will beef up Dell's infrastructure offerings, while VMware, which has a strong NFV software line, will operate as an independent satellite company, as it has under EMC ownership. (See Dell-EMC-VMware Merger Could Push Comms to Kids' Table and Dell Buys EMC for $67B in Biggest Tech Deal Ever.)
Which isn't to say that Dell has abandoned NFV. We'll talk more about that soon.
NFV leaders are still pursuing the vision articulated by CloudNFV. For example, HP Inc. (NYSE: HPQ), Heavy Reading analyst Roz Roseboro and AT&T Inc. (NYSE: T) distinguished engineer Margaret Chiosi have in recent conversations with Light Reading articulated a multi-stage process toward achieving NFV's full potential:
- Separating the software layer from the hardware layer, while still using a single-vendor solution for both.
- Virtualizing NFV software to run on a commodity appliance.
- Running NFV in the cloud, wherever it most makes sense from a technology and business perspective.
- Breaking up NFV from monolithic software into microservices that run dispersed over the network for optimal usefulness and performance.
Cloudifying NFV is an ambitious goal even today, and was, apparently, too ambitious back in the misty, prehistoric era of 2013. An aggressive, Dell-led CloudNFV is one of the intriguing might-have-beens of the comms industry. Instead, it's barely a footnote.