How APAC enterprises are using VMware Cloud on AWS

In a widely anticipated move, VMware finally made its VMware Cloud on AWS service available in Asia-Pacific (APAC) in late 2018, enabling enterprises running the vSphere-based technology stack to harness AWS’s infrastructure to power their applications, expand their datacentre footprint and facilitate disaster recovery.

Since then, VMware has nabbed several notable customers, including Australia’s ME Bank and, more recently, Singapore’s UOB Bank, which is using VMware Cloud on AWS to deploy and develop applications in a hybrid cloud environment.

In an interview with Computer Weekly on his visit to Singapore, Kit Colbert, vice-president and chief technology officer at VMware’s cloud platform business unit, offers insights on what APAC enterprises are doing with VMware Cloud on AWS, their deployment experience with the service, and VMware’s partnerships with other cloud providers.

What is the typical deployment process for businesses that are looking to move to VMware Cloud on AWS?

Kit Colbert: With VMware Cloud on AWS, we have greatly simplified the cloud migration process into a few simple steps.

First, we have condensed the process for setting up a software-defined datacentre (SDDC) into an application programming interface (API) call button in its user interface (UI). The SDDC set-up process is now fully automated and customers can do so in a very short period of time.

We also provide a number of tools that are part of VMware Cloud on AWS – Hybrid Cloud Extension (HCX), for example, is an all-in-one solution for workload mobility. If customers have vSphere on their on-premise data centre facility, we can actually move their workloads into the cloud with HCX and a few clicks of the button.

In terms of deployment, customers usually start with a trial of a couple of test apps on a smaller scale to see how things work. Once they are satisfied with the trial results and decide to commit to a full-scale migration, we then conduct a consultation session, and take them through the planning process for such a migration.

In the past, users would have had to contend with an arduous and lengthy process of individually reconfiguring every application and workload for the big move. Now, with a unified UI that combines many of the easy-to-use automated tools we have incorporated into VMware Cloud on AWS, they can not only enjoy a seamless UI that looks and feels like the on-premise environment they are familiar with, but also accommodate any custom-written scripts or third-party tools installed.

The automated HCX tool has also enabled us to greatly shrink the migration timeline from what could be a three-year process into one to two weeks. We have seen many customers moving dozens to hundreds of virtual machines per day with this tool.

In other deployments, we have customers who want to make the migration to the cloud, but also keep some of the workloads on-premise, for a hybrid architecture. In such instances, success hinges on the ability for us to ensure continuous delivery while integrating the customer experience with the migration process.

Again, this is well-integrated with the tools that we have on VMware Cloud on AWS. Not only are we able to support migration from on-premise environments to the cloud, but also from cloud-native environments to a hybrid one, so it goes both ways.

In terms of pricing, customers can also choose to start off with a small number of hosts, and scale up according to their needs, so that they are only paying for what they use. So, it’s cloud economics, but applied to the same kind of SDDC environment that they have on-premise.

What types of workloads are being deployed on VMware Cloud on AWS?

Colbert: Different clouds offer different capabilities and optimisations, and there might also be clouds that are more viable in certain regions and locations. VMware’s focus is helping customers solve the issue of system complexity and make everything as seamless and convenient as possible.

In the case of VMware Cloud on AWS, application programming interfaces (APIs) for AWS are naturally quite different from those for Microsoft Azure. Our hybrid cloud strategy then is to layer our software stack on top of these different environments, thereby achieving a consistent virtualised layer for the user, no matter the underlying infrastructure.

While it is very tempting to imagine a world where all applications exist in the cloud and customers don’t have to deal with the job of managing, operating and securing an on-premise datacentre, the reality is not that simple

Everything appears the same to the user, and they can move their workloads around without the need to reformat them. All the operational tools work in the same way because they all use the same vSphere and SDDC APIs.

The real drive here is to provide consistency for the user, so that almost any type of workload can be deployed to the cloud. In the case of VMware Cloud on AWS, we provide customers with a standardised management layer, giving them access to core automation products that help with governance across different environments.

We also have security products, cost management products, as well as tools like VMware CloudHealth. So again, we are not trying to change the underlying infrastructure, but to provide users with the necessary tools that will enable them to move their workload across different environments in a multi-cloud world.

In the fullness of time, AWS believes all workloads will eventually move to the public cloud. What is VMware’s perspective on this?

Colbert: Our point of view at VMware is that we believe the future will be very much a hybrid one. Wherever the customers want to move their apps, we will be there for them.

While it is very tempting to imagine a world where all applications exist in the cloud and customers don’t have to deal with the job of managing, operating and securing an on-premise datacentre, the reality is not that simple. We do think that in the future, we will be seeing applications everywhere because of the following three laws:

First, the law of physics. The law of physics states that over cable, light travels only at a certain speed and that dictates latency. In certain use cases, especially the internet of things (IoT), automated vehicles or Industry 4.0 and factory-type use cases, latency is very important. Latency of several milliseconds can sometimes result in drastic consequences – so transmission to the cloud can actually impede these use cases.

Second, the law of economics. With all these IoT devices, video cameras and analytics, you could send all of the raw data up to the cloud – but does it really make sense economically? Not only do you have to pay to send such large amounts of data, but the time needed to transmit the data is unjustified most of the time.

Take the example of surveillance video – you have a set of video data, and you wish to identify the person in the video. Do you send the raw video feed of several gigabytes up to the cloud, or do you make the analysis on-premise and just send across the insights garnered from a quick analysis? In a lot of use cases, it just makes more sense to do the analysis on-premise simply because it is the more economical way to do it.

Third, the law of the land. The idea here is that you may have regulatory compliance issues that necessitate data staying in a certain location. Now obviously here in Singapore, we are lucky because there is an AWS region here, but not everywhere is the same. Many European countries do not have mega-cloud presence in their areas, so they are going to need local servers in some shape or form. Some jurisdictions such as Indonesia also require certain sensitive data to be stored in domestic servers rather than foreign locations.

So, these are the three big reasons why we think customers will still need to maintain certain workloads on-premise. Our view is that there can be workloads anywhere – in the public cloud; in the datacentre; at the edge; retail stores; factories, etc. Our goal is how to make it as easy as possible to manage across all those locations, to manage at scale. And VMware Cloud on AWS has enabled us to do all of these for our customers.

The availability roadmap of VMware Cloud on AWS in APAC appears to start with mature markets like Australia, Japan and Singapore where cloud adoption is more mature and where AWS regions are available. Is that the general approach that VMware is taking to decide which markets should come next?

Colbert: Our general approach has been to start off in a small number of markets, to try to improve on the technology, and then expand. By the end of this calendar year, we should be in pretty much every AWS region in the world. We spent the first year or so proving up the technology, really making sure that we are mature in our operational capabilities for both ourselves and AWS.

Singapore was the third availability zone for VMware Cloud on AWS in APAC after Sydney and Tokyo, and the hybrid cloud service is now available across 13 AWS regions globally.

In the later parts of this year, we are looking at rolling out this service to other APAC markets such as Mumbai, Hong Kong, Osaka and Seoul. Of course, locations in other parts of the world include Sao Paolo, Bahrain and Sweden.

VMware said it jointly engineered the service with AWS in UOB’s case. What exactly was jointly engineered?

Colbert: UOB is the first organisation in Southeast Asia to enhance its portfolio of enterprise solutions with VMware Cloud on AWS.

Jointly engineered by VMware and AWS in compliance with UOB’s strict security requirements, the platform enables applications deployed in this environment to benefit from robust disaster protection and optimised access to AWS’s wide range of services.

It also reduces the cost and effort associated with migrating applications to the cloud by delivering infrastructure and operations that are consistent with those deployed within customer datacentres, and extending tools, processes and practices proven to support the most demanding applications.

Is VMware looking to partner with other cloud providers on a similar service?

Colbert: First and foremost, directionally, we do want to deliver VMware Cloud everywhere.

Apart from our partnership with AWS, VMware also has similar partnerships with leading cloud providers such as IBM Cloud, Alibaba Cloud and our partnership with Google Cloud was just announced at Google Cloud Next’19. Customers will be able to run distributed, cloud-native applications across Google Cloud and on-premises environments via its integration with VMware NSX Service Mesh.

That being said, we do have a special partnership with AWS and we are doing some really innovative things there. Our focus is really on fulfilling that vision on AWS and there’s a lot of work to be done.

Another point is that apart from the full SDDC stack, vSphere and so on, there are a lot of other technologies we have as well, for which we are collaborating closely with our various partners.

We are working with Google on things like our VMware PKS offerings, around our vRealize automation offerings, and also our SD-WAN offering, among others. With Azure, we are doing a lot around our virtual desktop offering Horizon. So, there are a lot of projects that we are doing from a partner standpoint with different mega-cloud providers.

Globally, we work with over 4,000 partners as part of our VMware cloud provider programme. In the long run, we believe that there is no one-size-fits-all solution; we want to be wherever our customers are and support them in whichever environment they choose to work in, so we are always open to working with different partners for the benefit of the end-user.