From the blog.

Managing Digital Racket
The more I tune out, the less I miss it. But that has presented me with some complex choices for a nuanced approach to curb
Complexity – My Friend, My Enemy
Over my years of network engineering, I've learned that the fewer features you can implement while still achieving a business goal, the better. Why? Fewer

Missing Synergies & HP’s SDN

1,313 Words. Plan about 8 minute(s) to read this.

As someone who’s been monitoring HP’s SDN strategy for years now, news that Bethany Mayer is headed to Ixia is rather interesting. Despite HP’s networking division having had some successes and gaining small bits of market share here and there, the fact they they are leaders in the SDN space seems to go unnoticed by the broader market. My speculation as as to why that is, exactly, is along the lines of missing synergies.

First of all, let’s take a look at HP’s SDN assets. I won’t detail all of them – just the most important ones, in my opinion. There are three.

  1. An army of installed switches that can run various flavors of OpenFlow.
  2. The VAN controller that can make those (and other OpenFlow) switches do something interesting.
  3. A group of partners that are writing applications to do those interesting things.

By “interesting”, I mean that the HP SDN applications running on top of VAN do genuinely novel things. Take Guardicore, for instance. Identify hosts that appears to be scanning the network, and then use OpenFlow to redirect their scans to an active honeypot. This approach doesn’t wait for the honeypot to be discovered, but rather goes out and finds conversations that would be engaged by a honeypot and sends them there. From there, traffic can be analyzed and the threat mitigated. Interesting? You bet. Not only interesting, but compelling – it’s technology every organization can use. And that’s just one example. HP has more applications, capitalizing on the fairly obvious (even though no one else is doing it yet) idea to build an SDN app store supplying apps that work with the VAN controller. HP can demonstrate genuinely winning, distinct technologies with a real value proposition for a wide array of potential consumers.

So, why isn’t HP networking winning more deals?

HP-Networking-Logo1This is where Mayer & HP’s organizational structure comes in, and I don’t have the answers. I can only play armchair quarterback here, but from the perspective of someone much closer to racking networking equipment than selling it, here’s my take.

  1. HP networking isn’t capitalizing on its server & storage channels. HP needs to start selling networking along with servers & storage as bundled deals, and push those deals aggressively. In accounts where I’ve been working with HP, networking has been a unique, off-to-the-side offering. HP can really tell a great networking story. They have the assets. They have SDN leadership with one of the most mature solutions on the market. Push it.
  2. HP sales doesn’t know how to push the networking solutions. I’ve had a few chats with the HP sales folks about their switching product lines, and the fact is that they just don’t understand enough about the product lines, what they are selling, what they are selling against (except generically “Cisco”), and can’t articulate the value the HP networking solution has to offer. An encounter like that leaves possible buyers with a lack of confidence that the solution can deliver, and that’s a non-starter. The Cisco name is so thoroughly trusted, that it is difficult to displace without a competing solution inspiring similar trust. Maybe that’s a myopic judgement based on my experience, but I have heard similar stories from industry peers. How about you? How well can HP networking sales articulate the value of the various platforms to you? Comments welcome below.
  3. HP still has a huge number of switches to choose from — too many. HP needs to trim down the model count, settling on a smaller number of products that go head-to-head with Cisco, feature-for-feature, footprint-for-footprint. Position these switches competitively, price to kill, and then sell SDN applications that enterprises will love on top. This is a notion to not “sell HP” as much as “sell against Cisco.” Customers can rattle off specs of the latest Cisco Catalyst and Nexus gear. Fine. Capitalize on that. Start with what they know as a jumping off point into what they don’t. Just get that model count down first.
  4. HP hasn’t integrated their security offerings into their networking offerings. The HP security business (ArcSight and TippingPoint most notably) is a BU that runs itself, and doesn’t see much synergy with networking as far as I can tell. In HP’s security BU, there’s a great IPS. There’s a serious firewall offering now. There’s a world class log event & dashboard solution. Why isn’t all of this hardware and software tightly coupled? Will VAN start pushing configuration into TippingPoint gear as part of business policy expression? AFAIK, there is no such integration (someone from HP tell me if I’m wrong), but that would be a fabulous thing to be able to demonstrate to a customer.
  5. HP isn’t capitalizing on the whitebox movement. Whitebox is, at least in these early stages, a different target market than the market I’m referring to in points 1 – 4, but it’s going to be a growing one that I personally believe will cross verticals and organizational sizes. I think Dell gets it. HP should get it, too. There’s good silicon and interesting form factors in the HP lineup. There’s hardly any hardware entrants in the whitebox space right now, so there’s much more room. HP can manufacture at volume. So, get in the space now and become a trusted supplier, using the Dell model to support the customer’s solution end-to-end. Might that fly in the face of the careful SDN ecosystem HP has built? Not necessarily. It depends on integration of non-HP NOS’s with HP’s VAN controller. If we assume OpenFlow is the great southbound equalizer, that shouldn’t be *too* difficult (says the armchair quarterback). In that context, integrating a wider group of vendors into the VAN fold is a win, not a lose.

Yeah, but what about the HP FlexNetworking stuff? Doesn’t that do a lot of what you’re whining about? Flex, as best as I can tell, is largely a branding exercise and not a technology. FlexWhatever is a way to group all of HP networking’s disparate parts onto a large diagram, but doesn’t address tight integration with servers, storage, or even security. Beyond that, I don’t think I can even call Flex a reference architecture. For example, you go to this page on FlexCampus, there’s exactly zero technical documents. As in, less than one. The same with FlexFabric. Okay, there’s this year and a half old poster. (And I’m sure there’s more documentation scattered about since some of it is in my Evernote database.) In fairness, perhaps I’m inadvertently pointing out the weakness of HP’s web site as a marketing tool with its “hacker black” colors, tiny fonts, ever-changing hostname URL, and gruesome inability to find much of anything helpful unless you have bloody-minded stubbornness.

With Mayer moving onto greener pastures that I’ve no doubt she’ll excel in, I’m wondering if the next head of HP networking can reach across the HP BUs to form the internal partnerships needed to really make the HP data center story come to life. Several robust pieces are there. Now it’s time to put them together. This is a story HP’s got to be able to tell in the not-too-distant future. I have a feeling Cisco’s going to be telling the fully-automated, software defined data center story with ACI, maybe not soon, but eventually. Oh, and VMworld is next week. You know that pesky little upstart VMware’s been telling the SDDC story for a while now. Hmm.

Your thoughts welcome below, especially to point out where my speculation is wrong. Happy to entertain alternate views and correct the stuff I’ve missed.