Å tÄpÃ¡n KÃ¡pl - Fotolia
While VMware has numerous engineers to produce and refine its products, it isn't shy about swallowing up companies to either bolster its portfolio or to infuse intellectual property into various offerings. Many of VMware's acquisitions of late have filled gaps in the company's software-defined vision.
As VMware has branched out from server virtualization to areas where it hasn't yet made a strong impact, the company has made a few purchases that perplexed many onlookers, such as when it picked up Zimbra in 2010 for its enterprise email capabilities. (VMware sold off Zimbra in 2013.)
Other purchases have been clear winners : Nicira, for network virtualization, and AirWatch, for end-user computing management, come to mind. Others have yet to prove themselves, such as when the company bought Desktone for its desktop as a service (DaaS) functionality. And still other acquisitions that didn't quite fit VMware's game plan -- SpringSource -- were handed over to former VMware CEO Paul Maritz as he left to head the Pivotal spinoff.
Which companies should VMware consider to fulfill its quest to build the software-defined data center and why? SearchVMware asked its advisory board members for their input.
I would argue that the gaps in the offerings are not through lack of acquisition, but perhaps too many products spread too thin.
The world is moving towards absolutely massive automated deployment. If you have any intervention in deploying several thousand guests a month, it means significant manual input into their creation. This equals a direct cost.
To this end, vCenter Orchestrator and vCloud Automation Center are the beginnings of full-tilt automation. However, based on my experience and discussions with users, these products are still not quite ready for prime time. Any company hoping for an automated solution out of the box will have issues. At the moment, there is a perception that these products are really "consultware," due to very little documentation [on them and a lack of] IT staff experienced in using them.
If you are one of the pioneer users, short of paying for consultation, you will find the documentation quite sparse. The further one moves from core vSphere, the fewer people are able to help, and also lack the required documentation.
While VMware's core business still greatly depends on the hypervisor, many of the acquisitions we have seen over the last couple of years focus on two areas: cloud and end-user computing.
VMware has invested quite a bit of money to boost its end-user computer portfolio. We've seen companies like AirWatch, Desktone and CloudVolumes all come under VMware's ownership and I don't think that trend is going to stop.
If we stop and look at where VMware might be lagging behind within end-user computing, one aspect sticks out: persona management. Although Horizon View contains a persona management piece, it doesn't seem to be as solid as other third-party providers on the market. From the View deployments I've looked at, persona management always seems to be the role that is filled by another vendor, not VMware. If I had to make a call on how VMware might fill this gap, I would probably look at companies like AppSense or Liquidware Labs. Provided they are integrated properly, I think they would be a great fit into VMware's end-user computing portfolio.
VMware has a huge track record of acquiring companies, but when you really look at the numbers, most of the acquisitions were done between 2010 and 2012.
Looking back at those years, VMware did not always seem to have a clear view of where the acquired technologies would fit into the company's portfolio. This seems to have changed because the Nicira purchase has led to NSX for network virtualization, and AirWatch fits into the product family for desktops and mobile management.
While VMware has made a lot of purchases, it still needs to find a way to integrate all the earlier products into its portfolio. It is clear VMware is working on it when you see the management updates in NSX. But for products such as vCenter Operations Manager, vCloud Networking and Security, and vCloud Automation Center, there is still work that needs to be done to make it worthwhile for administrators. Management Web applications are different for many products; the tools are lacking a common look and feel.
I really hope VMware will create a generic management and application development framework that can be used for current products and future products to be created or acquired.
First up is Liquidware Labs. VMware can't do what Liquidware Labs does, and they require that functionality, full stop. A related company to this is LoginVSI. VMware needs tools to test and certify their environments. LoginVSI is, put simply, the best of the best. Both Liquidware Labs and LoginVSI are tiny, bite-sized morsels today. VMware needs what they offer and right now, today, these companies offer that technology -- and the relevant engineers -- for a price far lower than growing that technology in house.
Next up is Veeam. With EMC under pressure to jettison VMware so that activist investors can get a quick cash injection, and the EMC federation under pressure from divisive internal politics, it's time for VMware to stop reselling a crippled version of EMC's Avamar. For the use cases of VMware's customers, Veeam is unquestionably the better choice. It would also buy them plenty of smart engineers. The question is whether Veeam has grown too big to be a serious consideration.
DataCore Software is also on my list. They make a virtual storage gateway that, if integrated into VMware's offerings, would allow VMware to truly commoditize array vendors. EMC sure wouldn't like it, but VMware shouldn't allow that to be a consideration. If the investment into VSAN is to mean anything, then VMware will have to play for all the marbles. This means no half measures: Go for the kill and make arrays truly "legacy." DataCore gives VMware the ability to do this.
A3Cube strikes me as an interesting startup worth VMware's time and money. Emilio Billi, who was trying to extend HyperTransport out of the server and make data centers truly modular, founded A3Cube. It is a very young startup trying to replicate much of that vision, but this time using PCI Express. I think the time is now for A3Cube. PCIe is in every server out there, and demands of today's flash are driving storage, distributed computing, HPC and even your average cluster beyond what more traditional interconnects are capable of delivering. Marrying A3Cube's technology and talented engineers with VMware's VSAN and marketing resources would result in the ability to stand up hyperconverged all-flash data centers that would make the gods themselves weep with envy.
What's more important than new acquisitions, however, is making use of the technologies VMware already has. VMware is sitting on a pile of great tech, but is hamstrung by the "thermonuclear wasteland" of its own internal politics. VMware could be the first vendor to deliver a "data center in a can" -- next generation converged infrastructure that delivers true software-defined infrastructure pre-configured as part of the out-of-box experience. They have almost all the pieces, but they lack the necessary leadership to seize upon such a bold vision, let alone earnestly attempt to implement it.
Pulling that rabbit out of the hat would take all the various fiefdoms of VMware working together with unprecedented cooperation. It would mean pivoting from a company that sells you tools to manage your infrastructure to one that provides a product that doesn't require those types of purchases.
It's also, inevitably, the play that every major vendor is going to make. Cisco, Dell and so forth haven't been acquiring all the pieces of the data center pie just to make a complete mess of vertically integrated stuff that still requires a small army of sysadmins to configure, manage and maintain. In this era of automation, orchestration and clouds, clinging to that opex-heavy model is madness.
So the question that needs to be asked isn't what should VMware buy, but what does VMware want to be when it grows up? The hypervisor is a commodity. Management tools are getting there, too, and they are being supplanted by automation and predictive orchestration. Ultimately, VMware is a feature, or it's a product. With the entire industry moving to cloud-style pre-canned consumption, the entire data center is the product of tomorrow.