When HP launched a network functions virtualization (NFV) group in February, we suspected it was meant to be big. Senior Vice President Bethany Mayer, who’d been running HP’s networking group, was picked to run the NFV effort, and the mix of servers and virtual functions that comprises NFV seemed like a combination HP would want to pursue.
But the initiative turns out to be big in a different way. Its reach will be companywide as HP tries to establish itself as an NFV leader, Mayer told us in an interview last week.
As part of that effort, HP has hired a name out of the carrier world: Prodip Sen, the Verizon executive who’s been chairing the NFV industry specifications group (ISG) at ETSI. While we don’t know the specifics about Sen’s role (we found out about his arrival only after the interview), he’ll clearly bring some real-world expertise to HP’s efforts.
In terms of products, HP sees itself offering four key pieces related to NFV. It’s got the base infrastructure (servers, storage, and networking), and it’s got cloud management capabilities in its Helion group of products, which includes HP’s own OpenStack distribution. On the orchestration front, HP offers its NFV Director. And finally, HP offers some applications — the virtualized network functions (VNFs) to be spun up by all these other pieces. Those include the home subscriber server (HSS) and home location register (HLR), elements related to the LTE evolved packet core (EPC).
Here’s our coverage of HP’s recent changes, including the appointment of Antonio Neri to replace Mayer:
- Prodip Sen, Verizon’s Former NFV Guy, Joins HP
- HP Gets Even More Serious About OpenDaylight
- HP Networking Names a New Head: Antonio Neri
- HP and Other Vendors Push NFV Ecosystems at MWC
- HP Starting an NFV Unit, Led by SVP Bethany Mayer
And here’s a portion of our conversation with Mayer.
I think a lot of people interpreted the NFV assignment to mean you’d been demoted.
Mayer: Nope. This is a huge opportunity for us. My performance speaks for itself as well — I grew that [Networking] business 12 quarters consecutively.
What’s the goal behind starting an NFV group?
Mayer: There is a huge upside opportunity for HP to disrupt and engage in the core network of a carrier market as a result of NFV. We really think carriers have a different mindset than they had just a few years ago around proprietary stacks and a hardware-based infrastructure where they had to build a new box to create a new service.
My group’s role is a pan-HP role, to marshall all of HP and direct them to this effort. That’s a combination of products. My organization funds product portfolios within the company to build products for the NFV proofs-of-concept [PoCs], and we organize the go-to-market around this opportunity. So, across the company, we have a committed team of carrier sales folks who are dedicated to NFV.
My organization is under the cloud business unit under Martin Fink. Their goal is driving carriers to the cloud, and my group is part of that.
So, how many people are involved with your group?
Mayer: It’s across the company — so, maybe 5,000 folks focused on this.
HP has announced more layoffs, so the law of average says you’ll be losing people …
Mayer: We’re basically staying very focused on the goal of NFV. Within the carrier organization, we are dedicated to that. [CEO] Meg [Whitman]’s decision to lay folks off, that’s an efficiency decision, which I applaud, actually. Those decisions about where the resources will come from aren’t in my purview at all, so I can’t really comment on that.
How will HP make money on NFV?
Mayer: We’re already making money on the PoCs.
On PoCs? Really?
Mayer: Some customers have spent money on buildling PoCs. That’s a very small amount, obviously. What we think is by the end of calendar 2015, moving into ’16, we should see a significant uptick in revenue, because what’s happening next calendar year is, we’re seeing more field trials emerge. We’re in one field trial with BT and we’re seeing others around the world, and in ’16 we’ll see a push from field trials into actual deployment. We see revenue opportunities from now through calendar year 2018.
Why should anyone use HP’s architecture for NFV, as opposed to someone else’s?
Mayer: First, it’s very open. Our goal is to enable a heterogeneous architecture for the carriers. It’s what they’re looking for. They’re not looking for a complete stack. They’re able to utilize our offering and our partners’, and we’ll certify it for them. If you talk to Cisco, they won’t offer you that.
In the cloud, we have built virtual private clouds and virtual public clouds. We have an OpenStack distribution — that’s unique to HP — and we bring a tremendous amount of expertise. And there’s our commitment to open standards and open-source. One of the keys to NFV is an open-standards-based environment and using open-source when it’s available.
Finally, I would say we are the market leader in terms of infrastructure. If you look at the hypervisors in the world today, they primarily land on our products.
HP has upped its OpenDaylight Project membership to platinum status. Would you say you’ve had a change of heart there?
Mayer: This isn’t a change of heart. I made the decision [to go platinum] before I completed the work at Networking. We wanted to have more of a voice to help guide OpenDaylight and continue to enable it to be a true meritocracy. We’ve contributed, and we think OpenDaylight has a lot of capabilities we could perhaps utilize. You’ll see us engage more using the OpenDaylight code base in some of our products. As we do that, we want to make sure we’re part of the board and are there to help make decisions.
OK, but why did your OpenDaylight participation start at the low, silver level then? Why not start platinum?
Mayer: We wanted to participate but not necessarily get to the level, at that time, where we would have to depend on the code base that was in OpenDaylight at that point in time. It was fairly immature; that’s why we moved down the path of building a controller ourselves.
We believed that because OpenDaylight was an open-source project, that a meritocracy was possible and would happen, so we wanted to remain fairly engaged, but it was early as far as utilizing any code or participating in that way.