In a couple of weeks I’ll be in San Francisco for VMware annual conference, VMworld. This will actually be my first time attending the even in person and I suspect I’ve chosen the best year to be there – there are some seismic shifts occurring in VMware’s business and I suspect VMworld will see some massive announcements to go with this. Here’s a round up of predictions, reflections on rumors and my ideas on where VMware should go. This post is a follow on from my previous post suggesting strategies for VMware to follow.
VMware to Enter the Public IaaS Space
Rumors have been doing the rounds that VMware is planning to launch a public cloud IaaS service, tentatively called Project Zephyr. The rumor goes that VMware has bought data center space in Nevada and will be setting up an operation that uses the vCenter Operations Management Suite, vCloud Director and Site Recovery Manager for failover and disaster recovery, along with with EMC storage gear and Cisco’s Unified Computing System (UCS) as the computing platform. Bear in mind that VMware joined the Open Compute Initiaitve (OCI), a Facebook initiated project that aims to increase the efficiency of data centers and open source the best practice for construction of data centers at scale. In the blog post announcing the move, VMware spun it as a move designed to help their service provider customers leverage the best design smarts – one thing VMware can not be accused of is being generous however and the OCI move has more to do with VMware using design smarts for its own direct gain than anything else.
Zephyr can be contrasted with VCE, the converged infrastructure JV between VMware, EMC and Cisco This is no partnership, this is a case of VMware using its in house solutions and, where necessary, third party tools, to build its own IaaS service. Bear in mind VMware’s recent purchase of Nicira which clearly threatens Cisco’s position. The previously happy relationship between the three VCE partners must have got a little strained of late. Vmware very much seems to be in open war against Cisco (or at least heading in that direction) – it’s hard to imagine that Cisco won’t wake up soon and fight back – what that means for VCE is unclear. The problem with a VMware IaaS play is that it is going to upset a host of different people. Not least of all are the existing VMware service provider customers who, are apparently feeling very uncomfortable about this move. VMware has been on the attack and sources have said that Zephyr is a direct response to service provider partner’s slowness to build cloud services – in response to that service providers have voiced their concerns at having to compete directly with the VMware mothership. I spoke with someone very much au fait with the inner workings of VMware who suggested that execution of an IaaS will require a big change in VMware DNA;
I just do not see anyone from the inside having the DNA to do that. They must have hired some folks with proper large scale operational execution. I have not heard rumors about that but cannot see them pulling it off with that. Otherwise, they have all the blocks: if they build directly on top of ESX without going through vCenter and vCloud, they have a great shot at it.
A Possible Legacy Business Spin Off?
The problems with VMware going public IaaS is that they have a salesforce that makes a massive proportion of its revenue by selling existing hypervisor products. These people are going to feel very concerned by an initiative that sees their bread and butter threatened by a disruptive product. In an insightful post surmising what VMware’s strategy may be, Simon Wardley suggested that an IPO may occur for VCE (or some other move that sees the existing business separate from the new, more commodity-driven, business. In making this prediction, Wardley points out that;
The reason being is that at this moment much of the buzz in financial analyst circles is around hybrid cloud environments consisting of a combination of private & public infrastructure and VMWare‘s technology is touted as well placed for this industry.
What the analysts don’t seem to understand is that this hybrid model is purely transitional and will be replaced in the majority by a hybrid model of public and public i.e. use of multiple public sources. The private enterprise cloud market is at best transitional and has a limited lifespan until it becomes niche.
In my post assessing what VMware should do, I suggested a standalone entity focused exclusively on cloud infrastructure, I further suggested this unit should include DynamicOps as a way of enabling a realistic and consistent hybrid Cloud play. Looking at my prediction alongside Wardley’s, I’m not convinced that his concept of a hybrid model of multiple cloud products will occur in the short to medium term. This is important because the timescale for a move away from public/private hybridization directly impacts upon the structure and timing for a spin out. In my scenario, the spun-out business is still a public/private hybrid provider but importantly has the structural separation from the existing business so as to avoid the problems of a sales force that seriously doesn’t want to sell the new products. Clearly the recent Nicira acquisition (which occurred after my previous post) should be lumped in with this business unit also. Granted it probably makes more sense to spin out the existing business since it is further along in its lifecycle and hence would attract a better level of investment.
I concur with Wardley’s view that right now AWS is artificially keeping pricing high so as to reduce demand. In a marketplace without significant competition that have the luxury of making super-normal profits while also having the time to scale what they do. VMware building out a cloud business to scale will create a price pressure that AWS will be forced to react to. As Wardley says;
I believe there is a reason why AWS doesn’t simply reduce its public pricing drastically and that’s because I suspect that AWS has a bottleneck – buying up enough land to build data centers to keep up with its rapidly expanding growth. Dropping prices aggressively will just increased demand and exacerbate this problem. Any price reductions will have to be carefully managed.
This is also why I think that the GCE (Google Compute Engine) focus on initiating a price war is smart because whenever you face a competitor with a bottleneck, creating a price war to increase demand beyond their ability to supply will naturally fragment the market.
Wardley suggests that VMware will launch a IaaS attack, powered by open source clouds (he predicts CloudStack). While I understand his reasons for suggesting that there is no way to compete on price while using proprietary software, I can’t help but think that VMware – who now owns almost every piece they need to power an IaaS offering – could build something using their own software – perhaps.
VMware Rolls Back into EMC
Another fascinating prospect that has been suggested to me is that EMC could roll VMware directly back into EMC. EMC has all the component pieces (or would do) and with former VMware CEO Pail Maritz now at EMC, he could arguably look at all the pieces across EMC/VMware to build a more powerful combination. VMware and EMC, despite the combined ownership, are poles apart in execution- Maritz could bring that together. In this model, VMware would fold into EMC thus ending the VCE partnership. Of course there would be something of a blood bath within VMware in terms of sales staff but reports from the street suggest that some bleeding is already occurring.
Where Does Cloud Foundry Sit?
In all of this the big unknown is VMware’s intentions for Cloud Foundry. I can’t believe that it is simply a ploy for selling more hypervisor level product. Rather I see VMware playing a game which sees them deliver a compelling proposition at both the infrastructure and the platform layer. They’ll monetize Cloud Foundry (hopefully in a way that avoids the destruction of the vibrant Cloud Foundry ecosystem) and then have a compelling proposition at both the IaaS and PaaS levels. Hence Cloud Foundry ideally sits with the new cloud entity as the public IaaS and PaaS provider but with good hooks into VMware’s existing products and hence legacy customers.
VMware is one smart organization – and my predictions come without having any briefings whatsoevers from the company. The changes I’m suggesting are pretty massive however and not something that will occur quickly. I see a six to 12 month timescale for the unwrapping of the new look entity (or entities) and thereafter the cloud market is going to get a whole lot more interesting.