HP’s startling announcement, two weeks ago, to discontinue Touchpad and all webOS-based products, purchase Autonomy Corporation, and split off its PC divisions, caught the market off-guard. Hewlett-Packard Chief Executive Officer Leo Apotheker feels the company could be the next Polaroid in the consumer products and mobile device war — a business that requires companies to be “much faster than a conglomerate can move in most circumstances.” The reality is this new strategic direction should not have surprised anyone who has read Leo’s résumé; it was the board’s intention to hire a strategic thinker who could evolve the company into a software and services organization by leveraging HP high-margin assets coupled with a few acquisitions. HP has one of the strongest orchestration software portfolios in the industry, which encapsulates everything from enhancing user experience through its APM solution all the way down to controlling Layer 2 through the Intelligent Management Center (IMC). With strategy toward creating and servicing cloud infrastructures, HP should examine what it has and figure out if its current networking portfolio differentiates the company, changes the way networking is done, and aligns HP’s networking division to HP’s strategic goals.

Three things I&O teams should think about when it comes to HP:

• HP Networking’s division has not done much for HP, which could be a red flag for I&O managers investing in their solutions.  HP could, at any time, could cut the networking division--like they did with their WebOS and mobile division--based on their razor thin results.  At the beginning of 2008, HP owned about 9.5% of the market and 3Com owned 10% of the market (port share data from Dell Oro report, 2011 Q1). Three years later, HP only owns 20%. Their growth in revenue market share can be attributed to the doubling of ASP for the ports from the 3Com line. Basically, 3Com was selling to the S of SMB, and now is selling to the new M market, which gives the product line a higher combined ASP or more market revenue share (15% YoY). For spending $2.7B (not including integration costs) on 3Com, HP has reaped 0.5% gain in port market share.
• If HP got serious with its “alternative” approach to networking, HP could step out from Cisco’s shadow and blaze a new trail; a new approach that infrastructure and operations managers are crying out for. Otherwise VMware won’t be spending so much money marginalizing the de factor networking solutions in the datacenter today. With its heritage and strong software portfolio, HP is in a better position than VMware to break down the technology silos that exist within operation and infrastructure teams and drive the efficiencies business are expecting from virtualization. With HP’s feature-rich software tools combined with HP Lab’s innovative work with OpenFlow community, they have ability to orchestrate an entire infrastructure full of abstracted resources.
• HP could sell off the networking hardware business and invest the estimated $5B (3Com + ProCurve value) HP would receive into developing a virtual network infrastructure (VNI) solution — which is Forrester’s vision for a Layer 2-7 network stack composed of virtual, abstracted software, not hardware. This could be done by building on their IMC capabilities, OpenFlow work, and HP Software Suite and acquiring technologies like the ones found at Zeus, Vyatta, BigSwitch, Nicera, and Aerohive to overlay on a x86 and generic commodity ASIC network. In a roundabout way, HP has been trying to commoditize networking by harping on acquisition and lifetime warranty costs; as Leo points out, there is no future for HP in commodity hardware. Thus they could leverage $5B from the proceeds of the hardware division and capitalize on the R&D money they would need to spend on switches, like the A10500, to catch up with the other networking vendors.

With HP’s partnerships, converged infrastructure, virtualized networking solutions, HP Labs OpenFlow innovation, and topnotch software orchestration tools, I&O professionals could create virtual network infrastructure that VMware could only dream to create, while aligning to HP’s software, services, and cloud strategy. HP’s VNI could offer I&O organizations a true alternative to:

• De facto networking architecture defined by Cisco
• Juniper’s QFabric
• Arista’s EOS
• Brocade’s ONE(SAN/LAN convergence)
• Avaya’s VENA(voice/video/LAN convergence)

The problem is HP had better get started soon. It’s only a matter of time before virtualized networking will be deployed. Already, the L4-L7 vendors have working solutions. As for L2-L3, current networking vendors (NEC, Extreme, etc.) are working with OpenFlow teams to create software controller-based solutions. New companies (Big Switch, Nicera) are being formed.

Do you think HP should try a different strategy? If they do come out with VNI solution, does HP still need the 20% hardware market share, mostly in the SMB space (based on ASP per port), to be considered a networking vendor? Does it matter who provides the commodity hardware if HP offers the management and control plane solution? If they do sell their hardware portion of networking, can they build network credibility off their services, ESS, and business software divisions?