Tweet This: Tweet
Share on LinkedIn:
By John Valentine, Kovarus Senior Enterprise Architect
Now that a few days have passed since 2018 AWS re:Invent, I have to say this year did not disappoint! AWS released so many new features (it seems like it doubles each year) that it was hard to keep track of them all. You’ll often notice trends in the keynotes and this year was no different. Many presenters spoke about data lakes, machine learning and hybrid cloud solutions.
Being an infrastructure guy, I tend to get excited around the topic of hybrid cloud. We saw a compelling keynote from AWS CEO Andy Jassy and VMware CEO Pat Gelsinger regarding that growing partnership and the many new features being offered with VMC on AWS. Many of you may have also seen the release of AWS Outpost, which is AWS’s debut into the data center, providing customers the same AWS hardware, software and support experience on premise as they get in AWS’s cloud. I’m sure we will learn much more about that over the upcoming months.
Where I want to focus is VMC on AWS. Prior to this year, I wasn’t quite sure what to think of the VMC on AWS solution, and I am pretty certain I wasn’t alone. The new features released this year have changed my mind, as it now seems like a very compelling solution for many use cases. What has the goal always been concerning VMC on AWS? The answer is seamless integration between your on-premises data center and the cloud. Unfortunately, that use case was limited to a very costly and rigid model that wasn’t feasible for many customers. However, there have been some compelling advancements and growth around the solution.
For starters, the number of customers using VMC on AWS has doubled every quarter since the launch of the solution and over 200 systems integrators and resellers (including Kovarus) have achieved competency in the solution just this year.
So, all that said, let’s look at the VMC on AWS updates and how they accomplish a truly complete hybrid-cloud solution, which should always address the following:
- Networking that stretches from on-premises to cloud with predictable high performance. Networking should allow for seamless application mobility and consistent network security. AWS has various methods for connecting to their cloud. This could be an IPSec VPN, which is a secure tunnel over a shared internet connection. For customers desiring better performance, the option to purchase an AWS Direct Connect link is also available, which provides a dedicated network connection with guaranteed performance. And lastly, customers now have the option of using NSX coupled with an ENI (Elastic Network Interface), providing high-bandwidth, low-latency connectivity to the customers AWS VPC. While network connectivity is an important part of the equation, it’s not the only piece that needs to be solved. Data transfer is always a challenge but with VMware’s HCX solution, many issues are resolved and allow compatibility between on-premises and the cloud. This provides the coveted ability to have seamless application mobility, which wasn’t so seamless and usually requires a full migration to AWS.
- Storage and compute that is elastic, global and scalable with cost-effective on-demand consumption. Originally, the unit of scale for VMC required using I3 instances and 15.2 TB of raw DAS NVMe storage per instance. While this fit many workloads, this proved to be too rigid for some customers, and in many cases limited VMC on AWS from being a financially viable solution. Hence the release of Elastic vSAN, which provides customers the freedom to scale storage and compute independently. This is backed by AWS EBS, allowing for large storage environments that mimic shared storage array and scale without needing to add compute. Simply start with a minimum three VM cluster and scale either compute or storage as needed.
- Management that provides consistent operational experience across both environments. This was always a strength of VMC on AWS. For customers that are familiar with the VMware tools such as vCenter, vSphere, vSAN and NSX will find the experience identical. Utilizing the public cloud without having to hire a full set of AWS engineers, or hire excellent consultants like Kovarus can save customers a lot of time and money. At the same time, customers can still utilize all of the AWS native services.
- And…most importantly, security that overarches everything and is consistent with established policies, processes, tools and governance. Customers can easily extend on-premises network security policies to AWS, as well as their NSX policies. And due to the nature of AWS, vulnerabilities such as DDoS attacks become much less of an issue. If needed, customers can also take advantage of AWS Nitro instances which provides security imbedded in infrastructure itself. Nitro local NVMe storage transparently encrypts all data using dedicated hardware and the Nitro security chip continuously monitors and protects hardware resources. If you’re interested in learning a bit more, check out this link — https://docs.aws.amazon.com/AWSEC2/latest/UserGuide/instance-types.html#ec2-nitro-instances
With a lot of these updates, VMC on AWS now addresses what I see as three main use cases below. AWS also feels that next generation applications are a use case, but I don’t necessarily see this as the case.
- Data Center Expansion — Because of the seamless management, networking and the newly released Elastic vSAN, customers can now feasibly use VMC on AWS for new projects that traditionally required hardware purchases. Many customers view this as a way to delay or cancel hardware refreshes in lieu of the cloud which ultimately drives down cost and replaces refresh requirement every 3–5 years.
- Disaster Recovery — Elastic vSAN and new networking features save the day here. Customers can now have a small footprint of three VMs in their VMC on AWS environment, but have as much replicated data as they need (sitting in EBS). In a DR situation, customers can take advantage of elastic DRS autoscaling, which provisions the needed hosts to match workloads, and then scales them down once the issue has passed. This, combined with the ability to vMotion hosts back and forth seamlessly, makes VMC on AWS a strong DR play. A customer win was PHH Mortgage, who moved 350 workloads from their existing DR colo into VMC on AWS to replace existing DR solution.
- Cloud Migration — This is a strong use case given that there is no real application re-tooling or rewriting that needs to happen. Since your workloads already sit in VMware, they don’t really care what hardware they live on. So, if you have a specific application you want to migrate to the cloud, you can easily do it. This also provides a great opportunity to avoid that looming infrastructure refresh every three to five years.
Many IT organizations are also facing either a cloud first, or a more challenging full data center migration to the cloud initiative. In this case, VMC on AWS becomes a very compelling story. I work with many customers that miss critical cloud migration deadlines due to unexpected migration issues such as networking, VM conversion, right sizing and talent deficit. Many of these are mitigated using VMC on AWS. MIT was able to migrate 2,800 workloads and 550 TB of data within 45 days to AWS, with the first 300 workloads being live migrated in less that 96 hours. Contrast this with the year or more it takes traditional migration efforts.
There are a few more updates I think are important —
- Stretched Clusters for HA — A new feature allowing customers to stretch VMware infrastructure across two AWS AZs. This includes stretched vSAN storage and stretched networking using NSX. This provides 4x9s workload resiliency guaranteed. This also requires no application changes to take advantage of it; the infrastructure is intelligent enough to handle the HA event. A good customer reference is Stagecoach, who maintains a 4x9s uptime running stretched clusters. Their applications require roughly 45 TB and track 8,000 busses every 30 seconds.
- VMC Availability — By the 4th quarter of 2019, VMC will be available in all the same locations as AWS Availability Zones. As a side note, one thing that will lower costs is to always position the VMware SDDC next to AWS VPC as this will minimize data transfer costs. Basically, this means you’ll want to use the same AZ for the VMW cloud as your own AWS VPC.
- AWS RDS on Premises — This was really cool! Customers can now run Amazon RDS on VMware, ultimately bringing one of the fastest growing AWS services on premises.
As you can see, AWS and VMware have done a lot to make VMC on AWS as easy to use as possible for their customers. If you’re looking for a simple solution for your next hybrid-cloud solution, this is definitely something to look at. While it’s a simple solution, working with a trusted partner can make the process much easier. At Kovarus, we are certified with VMC on AWS, are an AWS Advanced Partner, sit on VMware’s advisory council and are NSX Elite. We are happy to help!