Why Docker is the new VMware (part 2)

Last week a tweet by Duncan Epping based on my post in part 1 referred to a page from Massimo Re Ferre and I think Duncan’s intent was to downplay the effect of my post and emphasize the technological differences between the solutions VMware offers (with the bold mark on vSphere) and the new technologies that are emerging on the horizon (as Massimo mentions in his post and I highlighted before in the likes of Docker (or containerization of applications) and, even more abstracted, Lambda model as AWS does. The latter plays even more into the virtualization model I described in an even earlier post I wrote around 5 years ago (over here). (Talk about long term vision)

There is a distinct difference between the technological model and what actually goes on in board-rooms. There isn’t a single executive interested in which technology supports his/her business. The decisions being made over there are done in three way model. CAPEX/OPEX, Risk and ROI. These are mapped onto individual business requirements and the resulting technology is chosen from there. Currently VMware is extremely well cemented mainly because of the risk and ROI factor. (They are certainly not cheap.)

The reason for that is linked to history. VMware filled a niche market in the late 90’s early 2000’s which they were the only ones (besides the high end mainframe and UNIX vendors) who provides a method of creating virtual machines that just sit in software instead of having them bolted onto hardware. They were the first one of doing this on Intel based x86 processors platforms and this massively improved overall resource utilization which inherently greatly reduced the CAPEX and, because of easier management, OPEX. When HA features like vMotion and fail-over capabilities were introduced the risk-factor dramatically dropped allowing for a more wide-spread adoption in the data-centre. Over the years there hasn’t changed much in the concept of virtualization of operating systems and the applications tied to them. Of course VMware created a huge eco-system around their hypervisor and beefed it up significantly with all sorts of tools, features, gadgets etc etc. The switch from VMware server to a bare-bone hypervisor allowed for a significant reduction in host resource requirements which could then be better used by the virtual machines.

The inherent problem still is that there is a huge overhead in the necessity of needing to run virtual machines. If you have 400 VM’s loaded onto a host this still means you need 400 times the resources of memory, storage, CPU and the rest of hardware kit just to be able to run these OS’s. You haven’t even installed a single app or used a single CPU cycle on any business related transaction.

This brings me back to what is the essential back-bone of CxO decisions. You have to do more with less therefore reducing CAPEX/OPEX and improving ROI. The only thing holding back these decision points is risk. The maturity of technologies like Docker and AWS Lambda services is most likely the most important contributing factors of companies to not adopt these just yet but I guarantee you they are certainly on the radar and many companies are exploring them as I write this.

Duncan highlighted the point that it is not realistic to see applications change overnight and I agree.


That being said though, this is also the same argument that led to the demise of DEC, SCO, SUN, Silicon Graphics and others. Lack of diversity in revenue generating technologies (not products) made sure these companies do no longer exist or were easy targets for hostile takeovers. A similar thing that could have happened to Microsoft as well. Microsoft hammered on Linux 10 years ago about it (Linux) being the cancer of IT and should be eradicated . Come today and Windows in the data-centre is being bypassed by Linux left and right and the overall revenue from the Windows server editions has plummeted tremendously. The only reason Microsoft is still alive today is because there is not a real user-friendly alternative on the desktop and in the office application space hence they still hold around 98% market share which keeps them afloat (comfortably I must say. :-)). Azure, albeit a great platform, still suffers from adoption rates as there are many cloud-service providers out there. HP, Rackspace, AWS etc etc. and as such the competition is extremely high.

The danger for VMware is time. The limitation in diversity of the product-set all centralized around the hypervisor may be its Achilles heel. If VMware is not creating a technology set independent of ESX they may run into the same hole as Microsoft did with Windows Server. Companies who matter most to business related applications like SAP, Oracle and other CRM based systems are not tied to VMware or a certain OS. If they see a good business case in creating container based applications which they can sell/lease/whatever and increase their revenue streams they most certainly will. This will then totally bypass any need of a hypervisor, and accompanied VM/OS.

Obviously the ESX hypervisor eco-system is not VMware’s only revenue cow. The cloud services and alliances with other tech-companies make sure there is a fairly consistent amount of money flowing to Palo Alto but techno-companies are not known for their long-term commitments. If they see market opportunities which lead to more revenue streams and better profitability they most certainly will jump ship or at least start nibbling on the vSphere sales numbers.

To conclude. My earlier posts mentioned above were not targeted to try and differentiate between Docker and VMware on a technology level. I run around in this business for over 20 years so accusing me of doing that is an indication you did not catch my intention.

I have a massive respect for VMware, their people and what they have achieved over the last 15 odd years. Their product-set, features, functions and capabilities is second-to-none and the level of engineering, pre-sales, services and support is outstanding. If there is one company who I would trust in bringing in a new technology-set like Docker or AWS Lambda service it would be VMware. They should get moving though since time is ticking and there is no niche market anymore in this space which could allow them to set their product and technology platforms in an even stronger slab of concrete.

I hope this explains my previous posts and how they should be interpreted.



Print Friendly, PDF & Email

About Erwin van Londen

Master Technical Analyst at Hitachi Data Systems
General Info , , , ,