If Dell and EMC really do merge, expect massive, bloody consolidation
Cisco, Nutanix and NetApp all have plenty to worry about, too
EMC/Dell deal We're far from convinced that Dell acquiring and/or merging with EMC is a good idea, but if it did what kind of entity would we be looking at?
The crux of the matter is just how much overlap Dell is willing to endure. And the answer may be "lots". Dell's made much in recent months of giving customers choices. Exhibit A is the company's willingness to sell customers a conventional switch or a white box switch running a variety of operating systems. Exhibit B is its vast range of servers that come in just about every shape, size and configuration (and with ARM perhaps on the way).
But it's hard to see Dell enduring the considerable overlap in products it shares with EMC.
In storage, for example, EMC and Dell's storage arrays overlap in many, many markets. Choice is one thing. Servicing US$20bn or more of new debt is another and servicing it will make it hard for Dell to hang on to all the people required to keep competing products alive. EMC's R&D teams probably surpasses Dell's if only because they've been at it for longer and have attacked a wider range of problems. Over time, it would probably make sense for EMC's storage to dominate the combined range, although it could also be argued that Dell's Compellent-derived efforts carry less baggage and might be easier to build on. That EMC's range is a superset of Dell's should clinch the deal and ensure the Evil Machine Company leaves a legacy.
On the software front, EMC's big data and management software ranges make sense as part of an expanded Dell Software. So do EMC's backup efforts and RSA. Documentum? For Sale: one slightly soiled document management company. Pivotal could and should go two ways. Its consulting arm would give Dell's a jolt of cloud cred. Pivotal's software efforts will bolster Dell Software. GE's investment in Pivotal is a prize, not least because GE looks to be a mover and shaker in the internet of things which makes it a nice friend to have.
On servers, the entanglements are myriad and Dell would have all sorts of things to unpick. It's hard to see Cisco being happy for VCE to go on, because while it enjoys having another way to sell UCS, it would not enjoy nourishing Dell.
VCE's already found a way to sell servers from vendors other than Cisco with its VXRacks, which include SuperMicro servers. A Dell-owned VCE could probably pull off that trick again and include Cisco networking kit without eyebrows being raised. Dell's always keen on customer choice, especially with its bring-your-own-NOS switches, so surely offering a Cisco-connected VCE offering won't be a stretch. EMC's VSPEX range competes with Dell's converged server products. Dell partner Nutanix spends its days telling the world that VMware's EVO:RAIL hyperconverged appliances are doing it all wrong. Something's got to give.
EMC's done very well over the years by letting VMware operate independently, while also giving its efforts momentum by signing up for them on day one. EMC also kept VMware from being a direct competitor to its own products: how else to explain the arbitrary limits on VSAN scale?
EMC was able to pull off this trick because ESX and vSphere became de facto standards without its active help. Indeed, EMC just sat back and watched server consolidation drive demand for networked storage. Dell would be brave to do things differently, or in any way that brings VMware in so close that it becomes possible to conceive of Virtzilla as a pawn. Integrated stacks are flavour of the month, but only Oracle is bold enough to shoot for the closed ecosystem.
At a guess, Dell would let VMware continue pretty much as it does today. If, as is widely mooted, VMware is used as a tethered goat to draw out the value Dell needs to buy EMC, it could even be that VMware is set on a path to independence. If sensible, Dell would keep a decent VMware stake if only to keep its not-inconsiderable profits flowing into its own coffers.
Independence for VMware may not be a bad thing. Today, HP and IBM don't see EMC as a direct competitor because while its range is broad, it's not quite able to make a direct full-frontal assault thanks to the EMC Federation structure providing a fig leaf of independence for all members. If Dell were to make VMware an integrated part of its business, IBM and HP's affection for OpenStack would give both excellent reasons to advance it as an even more direct vSphere competitor. Letting VMware go might therefore be better for Dell by keeping it on the right side of the we-sell-everything line. It would also help Dell's Wyse business, and Citrix entanglements, if VMware were at arm's length.
Another outcome of a Dell/EMC deal would be to leave NetApp as the only pure-play storage vendor of note. IBM, HP, HDS and Oracle can all sell multiple parts of stacks. Standing alone might not be the worst place for NetApp to find itself. It may not, however, be a comfortable place from which to make a stand.
What about brand? EMC's a mighty one in enterprise IT. Dell ditching it entirely from day one would surely be foolish, although the gradual de-emphasis of an acquired brand never works well. Mike Dell's name is on the door and he's seldom shown much interest in sharing it with anyone. EMC could disappear. Or perhaps be warehoused for use in some future Dell venture.
On the management front, the EMC Federation has warehoused clever and experienced people who were worthy potential successors to Joe Tucci and equally worthy potential successors to Michael Dell. Whether Dell can keep technical luminaries aboard is another matter. The likes of NSX daddy Martin Casado at VMware and ZFS/DDSD men Jeff Bonwick, Bill Moore at EMC are proven innovators that a hundred rivals would love to hire. Dell will need to keep most such folk aboard.
Plenty of other office staff won't be so lucky. If Dell, as is widely reported, is borrowing tens of billions of dollars to pull off this deal it'll be hard to see the carpet for the pink slips in some offices. Plenty of offices will go because the combined company won't need multiple offices in many cities and will need to save cash in order to be sure of paying off the interest on its acquisition war chest.
Will the deal come off? When HP and EMC were talking about coming together, the chatter was less detailed than it's been in the last few days. The Reg's virtualisation desk expects something to emerge this week. For what it's worth, both EMC and Dell have given us the “we don't comment on rumours” spiel and VMware's carrying on as usual ahead of its vForum conference in Australia next week, complete with senior executives on offer for interviews. If those interviews are cancelled, it'll be a sign something's afoot. ®
Bootnote: Years ago, your correspondent heard a story to the effect that when EMC acquired Data General way back in 1999, it was found that DG's European entity was set up rather more efficiently than EMC's. Legal and financial contortions therefore saw EMC Europe folded into Data General so that the former company could take advantage of its prey's tax arrangements. So while EMC acquired Data General, Data General Europe first acquired EMC Europe. Or so the story goes ...