I’ve read that 80% of VMware customer only have vSphere and aren’t leveraging other tools and products from VMware. As a result, I’ve wanted to work on a white paper that focuses on alternatives to the VMware ecosystem. I simply don’t have time to work on the white paper and the scope just gets bigger and bigger. Instead of writing a static white paper, I decided to dedicate a page on the blog to a running blog post that will compare VMware solutions to other products on the market. It’s going to be a living document. By living document, I mean the content will change as I find new information or get corrected by the many vendors I’ll offend.

Alternatives to vSphere

I don’t believe there’s a realistic alternative to VMware vSphere for most enterprises. There have been some interesting attempts. Companies tried to make OpenStack a free vSphere. That’s not realistic. However, there are some interesting attempts at alternative commercial ecosystems. There’s plenty of data and comparisons out there for Hyper-V and KVM in general. I won’t waste time initially looking at those solutions. One solution of interest is Nutanix’s “new” hypervisor named Acropolis (names are harder and harder to come by). There’s a fairly decent introduction to Nutanix’ strategy over on ZDNet. Acropolis may not be solid competition to VMware vSphere today, but it’s something to keep an eye-on.

Eric Wright (@discoposse) posted a good overview on the considerations of multi-hypervisor environments. It’s a good start on understanding multi-hypervisor environments. I’ve found while there’s a lot of chatter around multi-hypervisor environments there’s very little interaction about the topic.

VSAN – Storage/Hyperconverged

There’s a lot of money in storage. Nutanix’s entire business started on the concept of replacing your storage array with commodity x86 hardware. Nutanix has marketed hard against VCE and the Vblock. From a technical perspective, I don’t understand the comparison. Unless customers were just looking at the wrong platform, Vblock and Nutanix don’t compete. I would never consider running the same mission-critical workloads positioned for Vblocks on commodity x86 storage. However, there’s a lot of profit in storage.

Continuing to use Nutanix as an example, a friend had an entry level 3-node Nutanix cluster quoted, and it was over $100K. Over a $100K for 3-commodity Intel servers and some IP equals some good margin. On the other side of The Valley is VSAN. To show the potential for profit VMware charges $1495 per CPU to enable an all Flash option within VSAN. Many bloggers/analysts note that there’s no real additional engineering to justify the massive uplift. It’s pure profit. That said, this is a crowded space and expect a good deal of content about vendors such as Maxta, SimpliVity, Scale ComputingSpringPath and of course Nutanix.

NSX – Networking

VMware NSX is a big bet by VMware. The acquisition of Nicira has destroyed their partnership with Cisco. I wrote an FUD busting piece on VMware vs. Cisco that I spent a good amount of time researching. Expect content discussing alternatives about OpenDaylight, Cisco ACI, OpenContrail and Pluribus Network to start.

Data Center Management & Orchestration 

VMware has a fight on it’s hands with OpenStack. While I’ve been both a fan and skeptic of OpenStack (it’s simply too big of a mission), it undoubtedly has some serious industry support. Everyone from EMC, IBM, Cisco, and HP are betting big on OpenStack. I see it as the main conduit for enterprise companies to battle both VMware and AWS. VMware has had a major focus on orchestration, I’ve written about how they want to rule the data center and will use OpenStack as a means to their end.

Outside of OpenStack, this isn’t exactly my wheelhouse. But, I’ll be looking at vendors such as VMTurbo in this section.

Containers

One of the things I like about VMware is that it’s a difficult company to sneak up on. VSAN and NSX show that they are willing to break a few eggs and partnerships to stay relevant. Even if they haven’t seen the adoption of solutions outside of vSphere, they try hard. Containers is another area that VMware has continued to expand their reach. VMware introduced their very own Linux distribution and container management strategy. Expect to see plenty of Docker and CoreOS in this section.

Comments Section

I haven’t decided how I’ll run the comments section of this page. I can imagine it can get unmanageable. I’ll let it ride open for now as it would still be searchable and may add value for some.

Revision History

6/15/15 – Cleaned up some errors in grammar and added link to @discoposse’s blog post.

8 thoughts on “VMware Competition

  • June 12, 2015 at 9:08 am
    Permalink

    Disclaimer: Nutanix employee here.

    Keith, I’m very surprised at your comment, “I would never consider running the same mission critical workloads positioned for Vblocks on commodity x86 storage.” Why do you think that proprietary arrays utilizing 1987-invented RAID, all kinds of complexity around HBA, zoning, masking, etc, an inability to scale, and the necessity of multiple manufacturer support somehow is more resilient/higher performing than Nutanix which utilizes the same webscale principles of all the leading cloud providers (you will never find Google, AWS or Facebook running Vblock)?

    Certainly customers don’t agree with you. We have the world’s largest customers – long-time Vblock users, moving their most mission-critical applications such as SAP, Oracle and Exchange over to Nutanix. Not only are they realizing both better performance and higher resiliency, but also a vastly simpler environment to maintain and scale as well as a much less expensive platform.

    As one of the earlier sellers of Vblock, I can give you first-hand assurance that a high price tag and big racks full of impressive-looking equipment does not equate to a better solution for mission-critical applications. On the contrary, as an IT administrator, the more mission-critical the app, the more I would insist that it run on Nutanix.

    Reply
    • June 12, 2015 at 9:13 am
      Permalink

      Thanks for the comment and the engagement Steve. As you know in very large enterprises. It’s much more than the technology – while some parts of the market may say it’s all about the app this just isn’t true. x86 based storage hasn’t proved itself in the largest of workloads for the industries and companies I engage which is mainly Fortune 500’s.

      While the typical VDI workload is a great use case for x86 based storage companies simply don’t trust x86 yet. This will change over time. The example that you guys gave of 10K mailboxes running on Nutanix is a start. There will be a tipping point and this page and my comment will actually change when we reach it.

      Reply
    • June 12, 2015 at 10:25 am
      Permalink

      Steve, you are absolutely correct about old RAID technology, but customers shouldn’t have to pay such a huge premium to get high performance and resiliency. I have always and still am a huge fan of Nutanix, but my number one complaint is the cost, and my second complaint is the lack of scaling. Now when I say lack of scaling, I am talking about compute vs capacity. I realize there is additional performance and resiliency when you have storage in every single node, but if I need much more compute power than I do storage capacity, I have to spend a lot of extra money just to get that compute power. At least with VSAN it is not required to have storage in every ESXi host. Not that I would use VSAN. Which is why I went a different route and purchased storage from Tintri.

      Tintri offers amazing performance, it was built from the ground up just like Nutanix, and does not use traditional technologies. One might argue that Nutanix offers additional resiliency because data is stored across several nodes, but that is only partially true. Let’s be honest when it comes to high end storage. Is it more likely that there will be file system corruption, or a cab failure? While a properly designed datacenter should not suffer from a cab failure, there is always user error. Meaning that a technician working in that cab makes a mistake and causes the failure. So if I own a NX-3000 series with three nodes and something happens to that cab, I am hard down. To resolve that problem I will need to purchase more compute / storage and put it in another cab or datacenter. Same thing I would do with Tintri to avoid a cab failure.

      My point is, I was quoted for a NX-3360 with three nodes. For the price it would cost, I can buy two Tintri T820’s, more compute power, and enable replication. By using Tintri for storage, I can easily scale out the compute resources by getting additional servers. Storage doesn’t scale out quite as well as I would like, but because they have such a good price point, I can more easily budget to purchase additional appliances.

      There are going to be different use cases and different size companies with vastly different budgets where Nutanix may make more sense. But in my opinion and in my use case, I am getting all the performance I need and will be able to easily scale in the future by going with Tintri.

      Reply
  • June 12, 2015 at 9:33 am
    Permalink

    When I left the large channel partner I used to work for over 2 years ago, I told my boss (the president) that if I ever lost a Nutanix sale to Vblock, that I’d quit the industry. Haven’t had to quit yet 🙂

    Reply
    • June 12, 2015 at 9:37 am
      Permalink

      Looking forward to those customers sharing their experience and lessons learnt.

      Reply
  • June 12, 2015 at 1:23 pm
    Permalink

    I read this and listened to the podcast… not one mention of Hyper-V from Microsoft? I know VMWare has a mature solid footprint, but some customers are realizing Hyper-V has matured recently.

    Reply

Leave a Reply

%d bloggers like this: