VMware is already the best (most competent) and most important (fastest growing and the source of the most innovation) system software company on the planet. But as successful as VMware has been to date, it is worthwhile to ask what lies ahead – and most importantly in what direction VMware is likely to go on some key business and technical issues. In order to understand the range of choices VMware has it is worth looking at both Microsoft and Oracle as points of reference.
The Oracle Strategy
Oracle’s business and technical strategy can be summed up in the following points:
- Own the high end of the market
- Have premium products that offer premium value and have significant product differentiation
- Lock customers into those products to the greatest degree possible with proprietary approaches to everything
- Be fundamentally hostile to third party partners – cooperate with no one. Partners are either to be killed or to be acquired.
- Structure the licensing so that the customer needs a lawyer to decipher the license agreement and try to extract the maximum amount of money out of each customer for every product and year of maintenance
The Microsoft Strategy
Microsoft’s strategy with respect to products and pricing is:
- Own the high volume, high market share and low price point in the market
- Have products that are good enough for 80% or 90% of the market and that appeal to broad segments of the market
- Be very open with API’s that allow others to work with your products
- Be very open to partners. If a partner can add value to a Microsoft product, then Microsoft will co-operate at a sales and marketing level with that partner
- Make licensing and pricing as simple as possible (you can argue about this one especially on the VDI front but in the global scheme of things VDI is a niche)
So Where does VMware Stand with vSphere?
vSphere is hands down the technical and market leading virtualization platform:
- vSphere is market share leader, and is also the clear favorite of enterprise customers with large scale virtualization projects and customers trying to virtualize demanding workloads. vSphere also has considerable traction in the SME and SMB markets, although in these markets vSphere faces strong competition from Microsoft Hyper-V which is especially appealing to Windows only accounts who have Windows only admin expertise. So VMware currently owns the high end of the market and is current the share leader, which is a difficult position to defend.
- vSphere is a truly outstanding product. VMware’s track record of providing major enhancements to vSphere on a timely basis with little to no show stopper bugs is unmatched in the system software industry (as a benchmark look at at frequently and well VMware releases major updates to vSphere in comparison to how infrequently Microsoft releases major updates to Windows (and how many service packs Microsoft needs to get it right)).
- VMware deserves enormous credit for rapidly evolving the functionality of vSphere and at the same time opening up that functionality to third parties via open API’s. This is an area where VMware has clearly learned a lesson from the success of Windows in this regard, and where VMware is fundamentally different than Oracle.
- However on the vSphere front, VMware is abjectly hostile to the very companies that leverage its API’s and partners with none of them. The standard joke is that if you make software VMware views you as a competitor. Unlike Microsoft who with for example, SCOM, cooperates at a technical, marketing and sales level with anyone who adds value to SCOM, VMware does not cooperate with any vendor who leverages its API’s at a sales and marketing level.
- On the licensing front, VMware made a highly controversial move last year when it added a VRAM licensing component to vSphere. This was viewed by many customers as a “tax upon success” as it had the effect of penalizing the companies that pushed virtualization with high levels of consolidation (the more workloads you consolidated onto one server the more VRAM you needed and the more likely you were to bump into a VRAM quota limit and have to buy another license).
VMware and the vCenter line of Management Products
VMware is also a market leading vendor of management solutions for its virtualization platform. Notable products here are vCenter Operations (to monitor and manage the virtual infrastructure), vCloud Director (to layer a private cloud on top of vSphere), and vShield (to secure the virtual environment). What is controversial here is that VMware has departed from its per CPU socket licensing model for vSphere and instead has chosen to license all management products on a per VM basis. This is not an issue for most management solutions as per VM is very much like per server in the physical world which is how people paid for management software for physical servers.
But the per VM licensing model is a problem for one product in particular and that is vCenter Operations. The problem comes from the fact that one of the things that vCenter Operations does for you is to help you manage your capacity. Part of managing capacity is understanding whether you can increase VM density on your servers or not. Let’s say you currently have 10 VM’s per server on 100 servers or 1000 VM’s. You decide to buy vCenter Operations Enterprise. Each 25 pack of VM licenses is $4,875. You have 1000 VM’s so you you need 40 25 packs of licenses which will cost you $195,000. After running the product you find out that you have headroom to add 5 more VM’s per server or a total of 500 more VM’s. Now you have to write another check to VMware for $97,500. This bothers many customers because the increase in density is something that they view as being their benefit of buying the product in the first place, and they resent having to pay another license fee to get that benefit. At the end of the day, this method of licensing vCenter Operations boils down to a density tax just like the VRAM pricing in vSphere.
So Where is VMware Going to End Up – as the Next Microsoft or the next Oracle?
For a company like VMware settling on the “right” pricing and licensing policies is an extremely complicated task. On the one hand there are tremendous pressures to maximize market share by capitalizing upon a first to market with a leading product advantage and to monetize that position later with follow on products. On the other hand it is simple fact of running a business that needs to hire extremely talented and expensive software developers (in a high cost of living area) that VMware needs to make money in order to fund the R&D that makes VMware into the product leader that they are today. It is also entirely plausible that since EMC owns the lions share of VMware, that EMC puts pressure on VMware to contribute profits to the EMC financial results.
This complicated set of factors makes it difficult for VMware to do what most companies it its shoes would probably do. Most companies in VMware’s position (there are not many that have ever been there – only perhaps Microsoft, Google, and perhaps the new generation of social media companies come to mind) would choose to optimize for market share, run the company at minimum levels of profitability now and seek to come back an monetize that market penetration later with add on products. Therefore VMware is probably not going to follow the pure market penetration path that Microsoft pursued in the early days of Windows 95/98/NT.
On the other hand, it is clear that the senior management team at VMware is way to smart and experienced to position VMware into a high end niche in the virtualization market in a way analogous to how Oracle has positioned itself in the database server market (Oracle is the revenue share leader in databases, Microsoft is the unit share leader in databases). The most perplexing Oracle like behavior at VMware has to be VMware’s hostility to the partners in its own ecosystem who spend money every year to exhibit at VMworld.
The long term key to understanding how VMware is likely to come down on this issue will likely come from the applications platform business unit (vFabric and Cloud Foundry). That business unit is run by Tod Neilsen who previously leveraged robust partner ecosystems in his successes at Microsoft and BEA. VMware clearly recognizes that any system software platform that has a durable long term position in the marketplace must offer API’s to developers of applications – and that applications developers of all stripes must embrace the platform. The applications platform group will have to behave as Mr. Neilsen did in has successful efforts at Microsoft and BEA (by aggressively partnering with anyone who embraced the platform) instead of how VMware currently behaves around the vSphere platform. Therefore the most likely outcome to the high price vs high market share question is that VMware will choose the route of market share and volume adoption by developers for Cloud Foundry which would allow the ex-Microsoft management team at VMware to repeat their previous success with Windows.
VMware prices and licenses its products today along a set of models that are not optimized for either pure market penetration (like Microsoft) or pure extraction of the maximum cash from each customer (like Oracle). These policies will likely ensure that VMware continues to dominate the high end of the market – especially in enterprise accounts, but that VMware will leave itself open to being eaten from below by Microsoft Hyper-V (especially in Windows only SMB/SME accounts). The long term answer to how VMware positions itself with respect to price and value in the marketplace will likely come from vFabric and Cloud Foundry, as both of these products are crucial to the long term strategic position of VMware in the market – and also will rely upon aggressive third party support to succeed.