VMware and the reversible cloud

VMworld-2013-San-FranciscoVMware has been a huge agent of change for enterprise computing in the past fifteen years of its life. As Quentin Hardy notes in The New York Times, its products were the catalyst that killed the traditional small-scale enterprise datacenter:

“VMware’s virtualization software … disrupted the old computer server business, and helped usher in the current model of big data centers and cloud computing.”

Yet despite its disruptive impact, VMware is very much the mainstream choice of enterprise IT when it comes to cloud computing — mainly because most enterprises don’t see themselves moving their computing to the public cloud. VMware provides the tools they need to consolidate their computing into virtualized datacenters and private clouds, avoiding a bigger leap into the unknown of trusting public cloud providers.

Unfortunately for VMware, this leaves it at a competitive disadvantage amongst the growing minority of enterprises that have made that leap into the public cloud. The big names in public cloud, most notably Amazon Web Services, don’t run on VMware’s products. If VMware is to remain the virtualization market leader as more and more enterprises follow this course, it needs to establish its credentials as an essential building block of public cloud platforms.

Public cloud doppelganger

Hence the opening day of its VMWorld annual event today saw the roll-out of its vCloud Hybrid Service, which provides a public cloud doppelganger for enterprise private clouds running on VMware. Although the substantive announcement turns out to have more texture than was originally expected, it’s still first and foremost designed (as its name makes plain) to fit into a hybrid cloud strategy.

The service is intended to be the prime choice of enterprises that want an on-ramp to the public cloud without having to port to a different platform — with the added benefit of providing an easy way back to private cloud if the public environment proves inhospitable. It’s more than hybrid: it’s the reversible cloud.

The extra engineering that VMware has had to put into vCloud Hybrid to make it competitive with established public cloud providers perhaps gives an indication of how much of a gap still exists between the performance level of the average enterprise private cloud and the likes of Amazon Web Services. The Register‘s Jack Clark a few days ago described “the gigantic task-automation system” codenamed Project Zombie that has cut the time it takes to do a cloud deployment from 72 hours down to just 90 minutes. VMware earlier this year injected $30 million funding into Puppet Labs, whose technology powers the automation. Clark explains the motivation:

Project Zombie was developed after VMware took a look at the major clouds of Microsoft, Rackspace, and Amazon, and realized that each of these companies has developed ‘their own flavor’ of advanced, proprietary technology to underpin the systems. Project Zombie is VMware’s own ‘special sauce,’ [VMware's automation architect Nicholas Weaver] said, describing it as ‘an automation platform that can scale out and do amazing things.’

Software defined datacenter

In addition, VMware has at its disposal the software defined networking technology it bought when it acquired Nicira last year and which, along with storage virtualization, provides the foundation for its software defined datacenter (SDDC) strategy fleshed out today. Siren voices such as Gartner’s Chris Wolf have been calling on the company to launch another phase of disruption, once again reinventing the enterprise data center at the expense of established hardware vendors:

If VMware wants to really get the SDDC to take off, it needs to rediscover its inner rebellious teenager — the one that got it to where it is in the first place. Consider successful public cloud service providers such as AWS. Amazon’s stack places a premium on software and sees hardware as a commodity. Yet VMware is pushing a software-defined data center mostly on top of enterprise-grade hardware from its partners. How do you get to be cost competitive with AWS when you place a premium in the entire stack while Amazon only places a premium in software? You don’t.

Trouble is, VMware is trapped by its customers’ commitment to a reversible cloud strategy — that desire to have a path back, to be able to move their existing computing workloads at will into the public cloud and back out again, rather than reinventing enterprise computing to really take advantage of the cloud environment.

While I’m all in favor of taking as much advantage as you can of cloud technologies to wring out valuable cost and efficiency gains from existing IT infrastructure, this is a short-term improvement tactic. There are two important reasons why it doesn’t deliver long-term strategic value — one comes down to hardware, the other to software.

Optimizing for cloud

On the hardware side of the equation, the VMware approach, as Wolf explains, is tied to traditional enterprise hardware platforms. This is comfortable, familiar territory for most enterprises but it’s not what the big public cloud operators are putting into their datacenters.

In order to optimize server designs for the high-load, power-efficient requirements of their cloud datacenters, they’re going to custom-built machines in such volumes that Taiwanese commodity server builder Quanta is reckoned to account for as mamy as one in seven of all servers shipped. Facebook is among those leading the development of the Open Compute specification that threatens to strip away any proprietary advantage that high-end server makers Dell, HP and Cisco may claim for their more expensive designs.

On the software side, the divide is even starker. So long as enterprises stick with traditional computing stacks they may well find that they achieve better performance by remaining on conventional server designs and resisting the move to software defined networks and datacenter infrastructures.

Different dynamic

But when software is architected from the bottom up to operate as a public cloud service, a different dynamic takes hold. Instead of a vertical stack that’s tied to a tightly coupled platform architecture, it becomes possible to straddle multiple cloud platforms. Here’s an account I recently wrote of how Workday is moving to a multi-cloud infrastructure:

Under the covers, Workday is moving to an architecture that looks more and more like the Netflix model. Over several years, it has been evolving its architecture into a set of distributed services, initially by taking discrete functions such as payroll, analytics and integration out of the core object server and making them standalone services that are called as needed.

More recently, the vendor has begun porting a significant portion of the underlying database layer to Basho’s cloud-native Riak distributed key/value store. David Clarke, VP technology services, told analysts at last November’s Workday Rising conference that moving to the NoSQL database technology would allow Workday “to store many petabytes of data across multiple data centres.”

This capability in Riak to run in more than one datacenter brings greater fault tolerance and elasticity into the core architecture, bringing it a lot closer to what Netflix is doing and even further away from the classic, tightly coupled, vertically integrated enterprise computing stack.

By making it easier for its customers to move their existing computing to a public cloud platform, VMware is giving them new excuses to avoid truly grappling with how much they need to rearchitect their computing to take full advantage of the cloud. That in turn ties VMware to the laggards rather than keeping pace with the leading edge of innovation in cloud platforms.

For now, though, the larger market opportunity is among the laggards. VMware must pace its own evolution to that of its mainstream enterprise customers, leaving it forced to choose a middle path between the leading-edge innovation of the public cloud platforms and the more muted, reversible tactics of those customers.

Disclosure: Workday is a diginomica premium partner.

Image credit: Businessman walking to cloud © peshkova – Fotolia.com; VMWorld banner courtesy of VMware.

Phil Wainewright

Phil Wainewright

Phil Wainewright has been a thought leader in cloud computing as a blogger, analyst and consultant since 1998. As well as documenting the transformation of 21st century enterprises by digital technology, he is co-founder of industry advocacy group EuroCloud.
Phil Wainewright

@philww

web cloud SaaS expert, blogger @diginomica, LibDem, dad
Phil Wainewright
  • benkepes says:

    Good post Phil – VMware is standing on a burning platform that they’re all to aware of. The problem is (to extend a metaphor) there’s folks within the company that love the heat and keep dousing the structure with more gasoline. The existing revenue stream are, as I suggested in a post today, similar to kids on crack – super hard to give up.
    That said, Gelsinger and co know all this and are working hard to change – expect to see some real carnage in both the company and the ecosystem in the months ahead if they’re going to be successful.
    More here – http://diversity.net.nz/out-with-the-old-in-with-the-new-vmwares-new-reality-and-the-carnage-to-come/2013/08/27/