Last week Stu Miniman blogged that QLogic CEO H.K. Desai announced to Wall Street on its earnings call that it had begun revenue shipments of its converged network adapter (CNA) product to Oracle. SiliconAngle then followed up Stu’s blog with a post positing that the deal was an exclusive. I haven’t been able to confirm the exclusive but the guys at SiliconAngle are right more often than wrong on these things as they’re heavily plugged into the Silicon Valley insider scene. Normally firms like to dual source adapters but Sun was one firm that was prone to do exclusives and limit its supplier base—so I suspect SiliconAngle’s take is correct and Oracle is continuing that trend.
Oracle hasn’t announced anything on this and it likely won’t. The company doesn’t like to pump up its suppliers unless it serves Oracle and there’s no upside in letting QLogic promote the deal. But I suspect because the revenue could be material that QLogic had to disclose the deal to the street.
Late last year, Oracle announced a deal with partner Brocade which in typical Oracle fashion was completely self-serving and the timing staged (of course the catalyzing VCE announcement was also heavily hyped which forced Oracle’s hand). EMC had just announced the VCE coalition with VMware, Cisco and EMC and Oracle wanted to try and suck the air out of any competitive momentum VCE would confer to Cisco and VMware. It timed the announcement of the Brocade deal to be in lock step with the VCE release and identified several pre-tested configurations that included EMC V-MAX and CLARiiON systems. Very clever on Oracle’s part. Oracle was sending a message to the market that the VCE deal was no big deal…we (Oracle) can do that too and fill in any gaps we have in our portfolio with a well-timed press release.
Is Oracle a Serious Virtualization Play?
We haven’t heard much on the Oracle/Brocade partnership since the press release. Meanwhile Acadia has brought on Michael Capellas to head the joint venture and supposedly he’s hiring a huge number of people, including Todd Pavone who is an up-and-coming star at EMC. Pavone was EMC’s VP of Global Solutions which means he owned all the company’s reference architectures (aka Proven Solutions). He’s a player and supposedly will report directly to Capellas. Expect Pavone to build a boatload of proof points around VCE configurations that the company can show customers. There’s also a rumor that Acadia is changing its name to VCE to describe more accurately what it is. I expect VCE will continue to put pressure on Oracle, IBM, HP and NetApp to match the joint venture in the virtualization space; particularly with VMware. VCE will have to delicately balance the potential conflict with service companies like Accenture and CSC. At the same time, VMware and Cisco will have to dance with IBM, HP, NetApp and other partners. Interesting times as EMC claims it absolutely isn’t going after the services business like Dell (with the Perot acquisition).
Oracle in my view is poo pooing infrastructure as commodity plumbing that is readily available and emphasizing application value. Its strategy in the stack wars is to play up applications, business intelligence and the database and use Oracle VM as its answer to VMware. In the world of Oracle, why would anyone want to use VMware to virtualize apps – rather just use Oracle VM and everything will be tightly integrated. Never mind the fact that Oracle VM is way behind VMware in terms of function and market adoption – this is Oracle and the company can be extremely persuasive with customers.
And the new Oracle will tell customers that if you don’t want to buy all-Oracle – no problem – how about some middleware to go with that SAP installation, or some servers or do you need some open source Java software – we don’t care…and pigs fly.
This is what makes Oracle a serious virtualization play. It owns the most robust stack and the business. Oracle owns IP in applications, middleware, databases and now hardware with Sun, the company is absolutely a serious virtualizaton play. Acadia’s real challenge in my view is how to appeal to application owners and convince them that virtualization is actually a good thing for service levels. In the stack wars, as seen in this infographic we did back in May, Oracle has the richest stack in the business. It owns everything but networking which it feels comfortable not owning. In my view Oracle is smart on that front (as is IBM). Why try to slog it out with Cisco and HP when you can partner with networking companies like Brocade and QLogic to make applications run better.
The Inside Baseball on FCoE
One interesting story line here is that we’ve been hearing FCoE isn’t coming for a long while and then we hear the marketing flow from Vendor A (Brocade) that they smoke vendors B and C (QLogic and Emulex) in CNA performance and so on and so forth. So here’s the question. If Oracle and Brocade are so tight and Brocade has this killer ‘end-to-end’ strategy and Brocade’s CNA performance beats the competition in Oracle workloads, then how come QLogic just won an exclusive for CNAs with Oracle?
The answer is obvious – because Oracle thinks QLogic has better CNAs than the competition – that’s why it’s doing an exclusive with QLogic. Is this a true statement that can be backed up with facts? Yes – take a look at David Floyer’s independent analysis of Emulex and QLogic CNAs from this spring. Brocade’s not even on the radar in CNAs because that business is a two-horse race between QLogic and Emulex. Stu Miniman summed it up with this post laying out the competitive landscape in the adapter market. QLogic is the favorite in the two horse race and Emulex is the viable second source. Brocade is a switch player with virtually no presence in adapters.
The bottom line is that Brocade is the leader in switching and is getting squeezed from all sides. QLogic is the clear FC and FCoE time-to-market leader announcing $10M in CNA revenue this quarter. QLogic is trying to drive harder into switching, going after Brocade and breaking out of its adapter niche to create some distance from Emulex. The company has clearly been out-executing Emulex for the past two years. The question now is can QLogic become a true arms dealer to the data center? In my view the company needs more switching successes for me to say yes but the strategy is viable nonetheless with HP as an early supporter.
Emulex is a firm in transition that is hanging on to fat FC margins while at the same time having a “staring contest” with a goliath named Intel. Emulex really has no choice. It wasn’t going to win with an FC-only strategy and as a result had to jump on the Ethernet bandwagon. It chose not to sell itself to Broadcom for $764M – which ironically is about where the company’s market cap is today. So for Emulex it’s all about executing better to get its valuation up and demonstrating to investors that remaining independent was the right strategy. Its earnings are out on August 5th so we’ll have more data points then. Suffice to say that Emulex has a strong customer base and good relationships in FC that it will need to prove transferrable to Ethernet.
The Bigger Picture
Despite the fact that customers don’t want to disrupt existing operations, there is a big change coming to the data center. Technologies are converging around Ethernet and FCoE; And while FC has a sustained roadmap independent of Ethernet the technologies are clearly converging via FCoE. In my view it’s not a matter of if but when. Storage system vendors will tell you they don’t care about protocols but there are two realities which can’t be ignored: 1) FC is a safer bet than alternative technologies for mission critical workloads and 2) The big storage system vendors make more money on FC and FCoE. These factors will lead to FCoE adoption. As we’ve discussed in our research for quite some time, the FCoE adoption curve will slowly build momentum and then at some point rocket. Vendors and users that downplay its significance are taking a risk. Our advice to users has been consistent – get the technology in the labs and figure out its capabilities now while risks are low. When costs come down…switch.
The real issue in the data center is how organizations with diverse applications will support cloud-scale workloads. Should they outsource to cloud service providers or build private clouds internally? The choice depends on a number of factors but the reality is that for the next decade, internal IT will be tasked with providing services that are more cloud-like in terms of simplicity, ease of provisioning and costs. The single most important technology to support this trend is virtualization.
A key issue to watch will be the degree to which virtualization suppliers will support mission critical applications. Oracle, VMware, IBM, Microsoft, HP, Cisco, EMC and SAP are all critical players here and to date none has proven that virtualization is a safe bet for mission critical apps. VMware is working hard on this problem and there are a few early adopters, but most customers remain skeptical and entrenched. The tension here is virtualization is the only path to cloud-like services for internal IT but it scares application owners because it puts a layer in between the apps and the infrastructure that could cause service level disruptions (e.g. performance, security and management concerns). The flip side is CEOs are sick and tired of watching Google, Salesforce and Amazon simplify the lives of small and mid-sized businesses while their internal IT remains stove-piped and complex.
Something’s got to give and that something is convergence. And like it or not – it’s coming to a theater near you with FCoE as a critical component of the playbook.