vSphere on AWS – A Deal with the Devil?

Thanks to John Troyer‘s always excellent TechReckoning newsletter I ended up writing down some thoughts around vSphere on AWS – a topic I’ve been discussing and reading a lot about recently. While I can’t say for sure that what’s below is unique to me, I don’t think I’ve read this exact twist anywhere else. If you disagree, let me know in the comments — I won’t be offended. Let’s get to it…

Note: what’s below is mildly cleaned up from email back to John but not overly much. If anything isn’t clear, let me know – it was partially written in response to TechReckoning Dispatch, Vol. 3, Number 16, October 17, 2016.

Theory: vSphere on AWS is a deal with the devil. “Who is the devil?” is the main question.

  • Fact: if VMware could have put Amazon out of business, they’d have done it already. They’ve not been able to because Amazon offers quasi-decent Platform 2/2.5 capabilities but unmatched Platform 3 services (the competitive moat has alligators with rocket launchers on their backs) for “Cloud Native Applications” getting written today. Everyone believes applications are moving a direction that favors AWS – how fast is the open question (see the 451 group kerfuffle at VMworld). VMware believes it will be a while and in the meantime this can hugely drive NSX adoption (still something of a solution searching for a problem from what I’ve seen outside of security/microsegmentation).
  • Fact: if AWS could have put VMware out of business, they’d have done it. They’ve not been able to because refactoring apps for Platform 3/AWS style services takes time. The bottleneck IMHO here is developer quantity – The New Kingmakers hypothesis only goes so far if there’s just not enough folks with the right skill sets (why AWS wants their channel partners to be “born in the cloud” and HEAVILY focused on services).

So to me I read this as a play where AWS & VMware are both betting on the timeframe around refactoring applications to Platform 3 (provide me a better buzzword and I’ll happily use it – maybe I should just say Cloud Native Applications instead).

VMware is betting it will take longer than AWS thinks and in return they’ll get widespread NSX adoption plus preserve + build on being the dominant place for Platform 2/2.5 apps for the foreseeable future (I do also completely agree about Vmware+AWS concerns around AzureStack).

AWS is betting that applications will get refactored faster than VMware does and in the meantime why not 1) drive even more economy of scale + revenue, 2) let people get closer to the AWS services/API’s and remove some potential hindrances to consuming them (“wow…elastic DRS is cool…what if our applications could that that at an app-level? Oh…all these cool next gen services/APIS’s are in the same datacenter at LAN latency….wow!”), and 3) embrace VMware so tightly VMware may not be able to let go.

What if vSphere on AWS actually becomes dominant? Is that a bigger risk for AWS or VMware? Data still has weight and takes time to move.

Call it a deal with the devil…for both sides. 😉 Who’s the smarter devil though?

Side-note: while I’ve not searched for it, what’s also been interesting to me is that there’s been zero commentary on all the traditional storage criteria that usually drive 70%+ of virtual infrastructure cost (IOPs, latency, space, etc.).

What’s not obvious to me is VMware’s next play if vSphere on AWS and NSX dominate – that’s great but there’s got to be a play after that as I don’t think anyone thinks Platform 2/2.5 apps will be a sufficient TAM for more than 5-15 years (I’m sure some of the smarter folks than me at VMware have lots of thoughts on that front). Hopefully by then NSX and the oft-touted but rarely realized Automation+etc. plays of the vRealize Suite will actually amount to something.

What do you think? I’d love to hear your thoughts on this topic.

3 thoughts on “vSphere on AWS – A Deal with the Devil?

  1. You make some great points as always. Regarding VMware stickiness and legacy apps living in AWS, it’s not only NSX but also vSphere that will live on, since there won’t be the regular “should we renew this?” decision points whenever existing hardware gets too old.

    I think the IOPS and storage latency conversations haven’t gotten much discussion because this partnership isn’t actually a thing yet, it’s only an intent to do a thing. Once people can spin up a vSphere host in AWS and do a side by side load test comparison, we’ll start to see lots of articles about that aspect.

    Liked by 1 person

  2. Pingback: Newsletter: October 29, 2016 | Notes from MWhite

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s