I love the VMworld, as I do VMare. In the last few years VMworld has been “the IT show” if the infrastructure space is your thing. Still today, a large majority of storage and networking companies participate with a booth and some sort of news about integrations between their platforms and VMware products (although less than usual).
Something is changing and it is changing very quickly.
However, it is clear that something is changing and it is changing very quickly.
Dell-EMC deal
First of all we are still in the middle of the Dell/EMC acquisition. This obviously doesn’t help, it has been slowing down the whole organization and, even though this problem will be over in a couple of months, the re-organization process will take time and, as usual, not all key managers will be holding their position… Both Dell and EMC are very keen in reassuring investors and customers about the post-deal re-organization, products and support but, as we know, there is always an ocean between saying and doing.
The Golden goose… but for how long?
Don’t get me wrong, Vmware is still the golden goose but the competition has changed and moving quite fast now. All the hype is no longer about virtualization, it’s all about cloud (private, public and hybrid) and containers.
Today VMware announced what we were expecting with two pieces of news about Integrated containers (VIC) and a new version of its Integrated openstack (VIO). But that’s it, and it is not enough to spike my excitement and more importantly the excitement of the end users /nor that of the end users who are already looking at these two technologies…
Partnering with a cloud loser like IBM looks very tactical and not a great long term strategy.
And again, even though Vmware started some sort of collaboration with Google on the public cloud front, Amazon and Microsoft have much more interesting stories to tell… stories that customers actually like. And again, partnering with a cloud loser like IBM (from the perception POV at least) looks very tactical and not a great long term strategy.
About VIC and VIO
While I have to say that the job they’ve done with VIC and its integration with VRA is admirable, even if you can spot all the flaws (due to its immaturity), there is still a basic unresolved issue.
In fact, the vast majority of end users who are adopting containers are doing so on Linux and they are looking at open source based platforms for the management (like, for example Kubernets). Most of them are not actually interested in putting containers into a VMware environment. Yes, it could be fine if you have 99% of VMs and a bunch of containers but why pay $5000-$10,000 per node in licenses to manage containers? It is much easier and cheaper to build a new linux-based infrastructure from scratch, isn’t it?
It is not unusual now when talking to end users, hearing VMware referred to as the “legacy platform” and containers or Openstack as the new or the future.
It is not unusual hearing VMware referred to as the “legacy platform” and containers or OpenStack as the new or the future.
The clouds
Microsoft, for example, has a complete solution ranging from Windows Server 2016 (with containers) down to Azure PaaS and SaaS. And it promises VMs as well as containers and data mobility.
Google and Amazon, can leverage technologies like Kubernetes to help build a multi-cloud strategy. If not today, It will be for tomorrow.
But where is VMware (or Dell or EMC for what it’s worth)? It certainly tried, but its public cloud efforts were ridiculous, with the wrong technology associated with bad execution. Always afraid to step on its partners toes, or to lose some license sales, or trying to please EMC by using its (wrong) technology. Now it’s too late, Amazon is winning the first phase of the cloud war hands down with Google and Microsoft as the only credible competitors at the moment. The recent Rackspace acquisition just confirms this scenario.
For example, think that all, and I mean all, VMware storage partners, including primary and secondary storage vendors. They are implementing solutions to leverage object storage tiering/archiving (on S3 APIs) or DR or else on GCP, AWS and Azure… one can only imagine how much money VMware (or EMC or Dell) is leaving on the table.
All analysts predict that the overall spending in smaller datacenters will shrink in the next few years in favor of larger datacenters and IaaS/SaaS. How will it be possible for VMware to stay competitive then?
All analysts predict that the overall spending in smaller datacenters will shrink in the next few years in favor of larger datacenters and IaaS/SaaS. How will it be possible for VMware to stay competitive then? The business will not experience a rapid decline, but it will be hard to see the same level of revenues and growth experienced up to now.
Closing the circle
VMware is still a strong company with a lot of faithful/loyal users. But the virtualization wave has run its course, while cloud and container technology need a different approach or, we could say, a different level of abstraction.
Containers and Cloud introduce concepts that are all about data and application mobility without caring about the infrastructure.
Virtualization was is all about efficiency and has changed the way we design datacenters, but the infrastructure as we know it is still there, and with applications relying on it! Containers and Cloud introduce concepts that are all about data and application mobility without caring about the infrastructure which becomes invisible (or transparent if you like) in the eyes of the applications. Perhaps one of the mistakes WMware is making is exactly this: they are looking at containers as an infrastructure technology while it is much closer to applications and developers than they would like to admit. An interesting video recorded yesterday at Tech Field Day shows very well Docker’s point of view about containers for example.
I do hope I’m totally wrong and that we will be seeing many more (and exciting) things coming from Vmware in the next few months (at VMworld Europe perhaps?!)
If you are interested in these topics, I’ll be presenting at next TECHunplugged conference in Amsterdam on 6/10/16 and in Chicago on 27/10/16. A one day event focused on cloud computing and IT infrastructure with an innovative formula combines a group of independent, insightful and well-recognized bloggers with disruptive technology vendors and end users who manage rich technology environments. Join us!
test
Enrico, I kinda disagree regarding the fact that users don’t want to run containers on vSphere etc.
Yes, there are cheaper alternatives (even WAY cheaper) if you look at the bare cost of the underlying infrastructure. But IMO you have to consider the hidden (or not so hidden costs) of running the whole thing. As Kit Colbert correctly highlighted today, you still need monitoring, capacity management, backups, etc.
In this context, the TCO might favor approaches like VIC that in fact leverage existing tools and processes. Even for greenfield scenarios, the impact of operations (even for next gen/cloud native apps) deserves a lot of attention.
PJ
Hi PJ,
Thanks for your comment.
I think that If you are planning to have only a bunch of containers in a vmware-based environments you are right. It doesn’t make sense to change the way you work.
But if your organization is seriously looking at containers as the technology to deploy the largest part of applications from now on, you’ll end up with a very large number of containers (probably many more containers than VMs over time) and it is much better to think at a new infrastructure specifically designed for that purpose.
Which (in our strategy) would be Photon Platform. Which, in your article, you did not mention.
Long story short. I didn’t mention Photon because Containers on top of Hypervisors does make sense only for VMware. Especially if you plan to manage everything with platforms like Kubernets.
*blue screen*
Don’t take it personally… but you aren’t frictionless. 😀
@Enrico, right. Still, the very nature of the apps you’re going to deploy on this container-oriented infrastructure matters and changes the rules of the game, one way or the other. If you’re into stateless (or mostly stateless) apps, you don’t need nor want complexity in the mix, which calls for ‘container native’ infrastructures. But! If you’re taking advantage of containers for ‘traditional’ app architectures, you still need a lot of the processes and tools that you have and need now. For the enterprise use cases, I’d say the amount of this specific kind of use cases is relevant.
Great article enrico I do feel the virtualisation time is over, last year VMworld was all about containers, apps, docker and photon and then I thought that there was a chance VMware was loosing its way, and maybe they still are, I see nsx, Airwatch and workspace one as the future they should follow, Nutanix and others are taking the HCI space so they do need to watch out.
Andy nash
Well, Mr. Signoretti is on to something with his argument that VMware is a legacy technology company. After all, VMware has been around since the turn of the century (21st century) and has dominated the market for server virtualization for many years. Nothing fails like success, and VMware failed to re-invent itself in response to the emergence of public cloud computing services in 2006. Does this mean VMware will disappear quickly? No, because over the past 18 years VMware has amassed a large customer base who will mostly stick with VMware in the short term. The problem for VMware is they will not be able to attract many new customers who want to buy into their legacy server virtualization technology.
Novell faced a similar situation after it dominated the market for file server software for 16 years. It eventually found itself fighting a rearguard action against Microsoft, and then Linux, even though it bought its own Linux distribution having also previously acquired the rights to Unix from Bell Labs. Novell made an effort to bolt new services on to NetWare by implementing a Java Virtual Machine. It gradually moved all of its NetWare services to SUSE Linux Enterprise Server, but nothing Novell did was able to stop the erosion of customers and revenue. Novell finally sold itself for $2.2B to Attachmate in 2011.
VMware is doing similar things to retain customers by bolting containers to their architecture and supporting their own OpenStack distribution. As Mr. Signoretti pointed out, VMware is not an application vendor. A significant factor in the demise of Novell was that it did not have an easy-to-use application development platform. Microsoft did. Linux did. Novell didn’t. VMware’s demise won’t happen tomorrow, but the seeds of its decline have already been sown.