Welcome from Day 2 here at VMWorld 2017 in Las Vegas.


This is just a quick and dirty run down of my opinions on the topic of vSphere on AWS.


Couple of links from VMware to be aware of, and that I used in putting this post together: 


Pricing Guide:  https://cloud.vmware.com/vmc-aws/pricing

VMWare White Paper: https://www.vmware.com/content/dam/digitalmarketing/vmware/en/pdf/cloud/VMW-TO-Cloud-on-AWS-USLET-101-MED-RES.pdf



What is it?

vSphere on AWS is exactly that.  It is a set of VMware SKUS, that runs the VMware Cloud Foundation Stack (VSAN/NSX/VSPHERE) on AWS hardware in the AWS region of the customer’s choosing.  Right now, this is currently only available in the US West - Oregon Region.  I haven't seen a timetable for rolling this out worldwide, but we are still early in this product lifecycle.

This is a now active product.  Welcome to week 1 of vSphere on AWS being generally available!  Remember that this is still only active in the Oregon - US West Region.


What’s the value prop?

-        This offering allows a customer to QUICKLY spin up a datacenter utilizing AWS hardware and VMware tools that IT already knows how to use.

-        This allows customers to take advantage of the AWS global footprint for geo expansion.  Obviously from a hosted vSphere standpoint this is future state, however access from vSphere on AWS to other AWS resources can be carried to any region via the AWS network.

-        Service can be completely on-demand, and allow the customer to only pay for what they need.  I'll cover cost more a bit further down, however the Reserved Instance model is still available as well.



-         Customer MUST start with a minimum of 4 hosts.  Once they have 4 hosts in a region, they can add a single host at a time from there, but 4 is the minimum.

-        Bandwidth between the customer datacenter and the AWS region is still an issue that needs to be dealt with. The cloud adds latency to transactions.  IPSEC VPN and Direct-Connect services are available here, but are additional monthly cost considerations.

-        NSX on the customer premises isn’t required, but greatly enhances the capabilities in the realms of DR, and automating failover.  Being able to stretch out to universal logical switches and routers, allow the utilization of the same IP space in use in the same VXLAN across sites. 



-        On Demand is ~$32 per hour.  This is 4 hosts (the minimum) up and running.  The customer can reduce that by about half by signing a three year reserved pricing agreement, or by about 1/3 by doing a 1 year agreement.  If they were to run the bare minimum as a permanently up AWS based datacenter, the customer is looking at $280K per year.   This pricing doesn’t include any data transfer, routable IP’s, etc.  This is purely from a compute perspective.  There are other AWS costs to be mindful of.


-        There is also a “Hybrid Licensing” credit available.  Details around this to reduce the costs haven’t been released yet, and probably will require some involvement from licensing folks from VMware.


Where does this fit?

-        Customers that are looking to get out of the hardware business, but still want to manage their legacy infrastructure workload with the VMware tools that they know and love.  This will be a great fit.


-        Customers looking to add a DR site that they can quickly spin up and manage with vCenter.  Data Replication and failover orchestration still needs to be addressed.  Customers spending this money, are probably looking at Zerto / Veeam with SRM, or Rubrik/Cohesity for that.


-        Customers that have a workload that fits well into the standard amount of CPU/RAM/Storage that are being offered as part of this sku will benefit well from this offering.