BACKGROUND IMAGE: iSTOCK/GETTY IMAGES
Virtualization giant VMware joined forces with the leading cloud provider, AWS, last year to unite their deep customer bases and corner the hybrid cloud market. Nearly a year after the companies unveiled their partnership, a few early customers have dug into the VMware Cloud on AWS service -- but most still seek answers on how it works, its costs and availability. Both companies have stuck to their initial vague projection of a summer 2017 release, but the tea leaves suggest an unveiling close to VMworld in late August.
Mark Lohmeyer, vice president of products for VMware's cloud platform business unit, spoke with SearchAWS.com about the partnership, which he says is still on track for summer availability, and what it means for each provider's customers.
What did your customers tell you that made this partnership with AWS necessary?
Mark Lohmeyer: Customers were saying, 'We want VMware as our leading private cloud partner and AWS as our leading public cloud partner, to work more closely together. Fundamentally, today, you're forcing us to make this binary decision that we don't want to have to make. Either we can have a great VMware environment, but it's difficult to get the full advantage of AWS, or vice versa.' They didn't want to have to make that binary decision. They wanted us to work more closely together and come up with a solution that would effectively help them get the best of both worlds.
Effectively what we're doing is enabling the complete VMware software stack -- vSphere, vSan and NSX across compute network and storage -- to be delivered as a cloud service on top of AWS elastic bare metal infrastructure. The bare metal piece means we're running ESX, our hypervisor, directly on bare metal hardware, so it's not a nested solution. All the capabilities that a customer today gets the benefit of with ESXi running on prem, they can get on top of this bare metal instance from AWS.
Equally important to the fact that [the service] is bare metal is also that it's elastic. You can ask for additional capacity through an API and that capacity shows up a few minutes later. You can pay for that capacity on an hourly on-demand basis and only pay for what you actually consume. We think that's a key part of the best-of-both-worlds equation, that it's our stack running on top of that elastic bare metal infrastructure from AWS. A big part of the joint development effort between the two companies was around making those two pieces come together.
The idea that [an administrator] can leverage [an IT team's] existing tools and technologies, and leverage all these great capabilities of VMware on prem and take advantage of them in a flexible cloud-like consumption model on top of AWS, which is for many of them their public cloud of choice -- that value resonates incredibly strongly. [Customers] see a path where they can take advantage of that investment that they're making on prem and take it into the public cloud in the future if they'd like. And when they do run workloads on VMware Cloud on AWS, it actually brings together more elements of our stack, because every time they're consuming workloads on top of VMware Cloud on AWS, they're always taking advantage of the full stack. It brings along with it our advanced capabilities across the portfolio.
What's different between the AWS partnership and the VMware-IBM cloud partnership?
Lohmeyer: In both cases we're enabling the capabilities of the VMware software-defined data center, vSphere, vCenter and NSX in conjunction with those partners. The main distinction between them is, VMware Cloud on AWS is a VMware-sold, supported and delivered service. So the customer purchases, consumes and gets support through VMware. We're working behind the scenes with AWS to power that overall service. In the IBM case, it's an IBM-sold, supported and delivered service, so the customer would work with IBM to purchase, consume and get support for that service.
What management capabilities will be available with the VMware-AWS integration?
Lohmeyer: There's a number of things we're doing from a management perspective. The first is that the service is based on vCenter as the management control plane. Out of the gate, we've got hundreds of thousands of IT administrators around the world that use vCenter to manage their on-prem environments today. They can leverage all the capabilities of vCenter as they exist today.
On top of that, we continue to augment our vSphere platform and vCenter itself to deliver new value in these hybrid cloud environments. One of the new capabilities we're delivering is called Hybrid Linked Mode that links multiple vCenter instances together and presents a common view as a single control plane. [A customer can] link your vCenter that's running in VMware Cloud on AWS to your vCenter that you might be running in your own data center. And then, when you go into the vCenter UI, you'll see all those resources in the left-hand pane: your data centers, your hosts, your clusters, your VMs. You'll see them all: your clusters running in your own data centers as well as your clusters running on VMware on AWS. You can basically operate in that environment exactly the same as if it was clusters running in your own data centers.
Mark LohmeyerVP of products for the cloud platform business unit, VMware
Another one is called Elastic DRS. Distributed Resource Scaler is a very popular feature for VMware customers on prem [to] rebalance workloads across a cluster of hosts, [and] optimize utilization of the underlying hardware while still providing great performance for the applications that run on top. But as you add more workloads into the cluster, at some point your cluster is going to run out of capacity and you'll have to [add] hardware and provision it. Likewise it's not easy to scale down that cluster either.
Elastic DRS will allow you to set a policy for your DRS cluster in VMware Cloud on AWS that says, 'When the utilization exceeds a certain threshold, automatically reach down into AWS, spin up another host, add it to the cluster and rebalance my workloads across that now-larger cluster.' And if the utilization falls below a certain threshold, give that host back and then rebalance across that smaller cluster, then you don't have to pay for that hardware anymore. It doesn't require any changes to [how] the application runs today. The customer just defines the need of the applications and then VMware automatically scales the capacity of AWS up or down dynamically based on the need of the workload. The customer is only paying for what they need at any point in time. It provides a lot of agility and TCO [total cost of ownership] benefits, and it doesn't require a refactor of the application into some new approach.
What feedback have you received so far from the beta program?
Lohmeyer: We have approximately 50 customers and partners that are participating in the beta -- we had over 1,000 customers and partners who raised their hands and said, 'We want to be a part of the beta.' One [capability] they really like [is] that we made it simple and fast to get started. You just come to the VMware Cloud on AWS console [and] say what you want. They really like that simplicity and ease of provisioning and configuration.
The second thing that really resonated was, as they got into it and started using it, the vCenter is the new H5 client. It's fast, it's modern, it's what they're used to. There's no learning curve. ... It's the same vCenter that they know and love.
The last big thing is we're [integrating] chat into the service itself, so the customer who is using the service can live chat with a VMware employee so a question can get answered quickly.
What architectural improvements enable the ease of use you described?
Lohmeyer: We've created a new web console that the customer can [use] to purchase and provision the service. It's brand new, built from the ground up as a cloud-native service. Customers [can] easily go through the process of picking which AWS region they want to run in, how many hosts would they like to consume, how would they like to consume it -- hourly, on demand or other consumption options.
And the work we've done between VMware and AWS in terms of enabling the VMware [software-defined data center] capabilities to run directly on their elastic bare metal hardware, that's what's actually happening behind the scenes. Let's say the customer wants to select four hosts in the US-West Oregon region from AWS, hourly, on demand, they would do that through the VMware on AWS console and then once they click 'Go,' that's where we would deploy the VMware software-defined data center on top of AWS' elastic bare metal hardware and spin that up very quickly for the customer.
When you say 'spin that up very quickly,' is it instantaneous, as it is with AWS instances?
Lohmeyer: Yes, the idea is to make it completely on demand like it is with AWS. In terms of our service, if it's a first-time new customer, the first host would take approximately two hours, which is much faster than they can do typically in their own data centers. But then, from there, adding or removing individual hosts, which we think will be the most common operation ... adding or removing a host will be a few minutes.
We've seen different general timelines for the VMware on AWS service, saying 'summer 2017' and 'mid-2017.' Can you clarify when customers can expect the service?
Lohmeyer: We're on track for summer of 2017. I think it's just semantics, different people using different semantics for the same [release timeline]. What we've been communicating pretty consistently, I think, is that summer of 2017 is what we've been shooting for and what we're on target for.
One recent report indicated that AWS is developing software for corporate data centers with VMware. Can you provide any details on any extension of the partnership, or the agreement that you have with AWS?
Lohmeyer: We're not commenting on any of those rumors or speculations.
Editor's note: AWS did not respond to an interview request regarding the VMware partnership.
Customers have questions about VMware-AWS partnership
Google acquires Nutanix in hybrid cloud push
AWS adapts its hybrid cloud mentality