Any App, Any Device, Any Cloud. Sound familiar? If you attended VMworld a year ago, you likely heard a similar message. Unlike many technology companies who change direction with the tide, VMware is staying true to its message from a year ago: delivering the same experience across devices, platforms and infrastructure.
The "Any Cloud" message sounds simple, doesn’t it? After all, the promise of Cloud is the "Easy Button" for infrastructure: provision, deploy, scale up, scale down - Just swipe your card. Unfortunately, many of the early adopters of Cloud, especially at scale, realized it's not that easy. How many screens does your infrastructure team have to burn through to provision a new server in your data center(s)? Now, compare that with how many screens they use to provision a new compute resource on AWS, Azure, or GCP. My guess is the workflows are completely different: different tools, different processes, different expertise. It's bad enough if you only use one Cloud provider; what if you use two or more for diversity?
I have to admit, when I heard the announcement almost a year ago from VMware that you could now provision ESXi on AWS, I was skeptical. Now I get it – and it's exciting. I see a number of real-world use cases:
Flex-out to the Cloud. This is one of those early Cloud promises that is hard to get right. If you're running an ESXi shop on-prem and want to move your workload to Public Cloud running some other hypervisor, how easy is that in reality? Now imagine your data center simply extended into AWS running dedicated, bare-metal servers with ESXi. Take it a step further and picture using your existing vCenter to manage the AWS instances just like you manage the on-prem data center. Want to vMotion a VM to the Cloud? Right-click -> Migrate. Even better, turn on Elastic DRS and let the platform do it for you, automatically. All your templates, automation, and monitoring extend into AWS. That's powerful, yet simple.
Data Center moves. This topic comes up a lot in conversations I have with clients. Moving data centers is very expensive. Most times, you don't want to pick up your physical infrastructure and move it because there's too much risk and not enough of a downtime window to do so. In those cases, customers often have to purchase the dreaded "swing gear" from their infrastructure partners like Dell/EMC, Cisco, IBM, etc. What if you could just push your virtualized workloads to AWS, decommission the servers/storage at your current data center, move it and then re-hydrate those assets from AWS when you're ready? The option to pay per-use makes this especially cost-effective because when you're done, you spin-down the AWS resources (if you want) or simply scale back the footprint.
Disaster Recovery. Instead of buying/leasing an out-of-region data center or leasing rack space from a colo, why not use the global presence of AWS? As the service expands from US-West across the US and then overseas, the possibilities are almost endless for placing workloads for DR in geographic areas to meet the needs of regulatory compliance and risk mitigation. Placing workloads in different regions or Availability Zones enhances the resiliency even further – all without having you or the client invest in all the infrastructure and overhead.
There are so many other benefits, it's hard to name them all. VMware and AWS have also done some interesting work in the networking side so that your VMware VPC can also use the native AWS services like EC2, S3, Dynamo, Lambda, etc. without incurring cross-VPC data charges. Now you can deploy cloud-native apps alongside the more traditional services and not worry about network connectivity.
If you attended VMworld, you can access breakout sessions on-demand post-conference. Be sure to check out Scripps Networks presentation on their journey to VMware Cloud on AWS, aided in large part by RoundTower Technologies: "Architecture and Tooling for the Brownfield Hybrid Cloud," LHC1158BU.
For more details on the announcement, visit https://cloud.vmware.com/vmc-aws.