Zerto Virtual Replication for Amazon AWS
By Shannon Snowden, Zerto’s Sr. Technical Marketing Architect
In another blog post , we show you the new user interface (UI) in the upcoming release of Zerto Virtual Replication (ZVR) 4.0. In this post, we will dive into our upcoming support for recovery to Amazon Web Services (AWS)
With ZVR 4.0, we are excited about Zerto adding support for Amazon AWS Disaster Recovery. This is the next step in the Zerto Cloud vision by adding the public cloud as a target recovery site.
The idea behind Zerto Cloud is to federate dissimilar environments for data protection and mobility, and keep a common user interface to make management easy. The underlying infrastructure becomes a secondary concern to normal operations, and the business can focus on Service Level Agreement (SLA) requirements for the workloads.
It’s one thing to create a product that replicates to AWS, but it’s another thing to meet the high standards of Zerto. When we decided to add AWS as a target, a foundational requirement was that it had to maintain the same protection model that our enterprise and cloud customers enjoy; it also shouldn’t feel any different when performing normal administrative functions. A key feature to meet our requirements is the Zerto Virtual Protection Group (VPG).
Virtual Protection Groups VPGs: The Foundation of Robust Protection
As evidenced by our happy customers, Zerto has set the bar really high when it comes to replication and orchestration of disaster recovery and migrations. VPGs are a key component to making it happen.
VPGs let you collect together VMs into a logical protection group, regardless of their server or storage location. The VMs in the VPG are then replicated with their relationships preserved.
This allows for an application-centric management approach where you identify all of the VMs that make an application function, put them into a VPG and they are protected together. Zerto takes that to another level by having write-order fidelity between the VMs in the VPG. It means that when you pick a point in time (PIT) as far back as five days prior, with consistency so that all of the VMs recover together at the same point in time.
ZVR to AWS Architecture
The architecture for ZVR to AWS essentially matches what is found in an enterprise ZVR deployment.
On the source side, we see a standard deployment using either Hyper-V or VMWare infrastructure. This site, or it could be multiple source sites, is connected by VPN. On the AWS side, Zerto uses an Amazon Instance called the Zerto Cloud Appliance (ZCA) to run the Zerto Virtual Manager (ZVM), the Virtual Replication Appliance (ZVRA) and the Zerto Virtual Backup Appliance (ZVBA).
Amazon S3 storage is used to house the replicated data as well as the Journal that enables point in time recoveries. Amazon S3 is also used via the ZBA for Zerto Offsite Backup for more about Offsite Backup, read here.
To keep the administrative user experience the same as our enterprise and cloud customers enjoy, Zerto retains the familiar simple installation requirements. For ZVR to AWS, it’s remarkably the same as an enterprise installation. This usually takes less than an hour to be up and running, and protecting workloads.
Here is a summary of installation steps.
- If you don’t have an Amazon account, you will need to create one. You can do that here: http://aws.amazon.com/
- Establish a VPN connection. AWS calls this a VPC. For more on VPC requirements, see http://aws.amazon.com/vpc/
- Create a Windows Server 2012 R2 Instance from one of the Amazon AMIs. For more on AMIs see http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/AMIs.html
- Install the Zerto Cloud Appliance. The Zerto Cloud Appliance is a Windows installation that adds all of the necessary Zerto services for the target site in AWS. During the ZCA installation, you will need your AWS Access Key ID and Secret Access Key.
- Pair the sites
- Create VPGs
After synchronization is complete, you are able to conduct failover tests, or failovers to AWS. When doing a failover test or failover, ZVR will create an instance that you manage in the Amazon console. To see this in action, watch this demo.
More Protection Options with ZVR to AWS
Over the years of tracking production environments running ZVR, an expected Recovery Point Objective (RPO) is usually measured in seconds. It’s not odd to see 5 or 6 second RPOs in a heavily used production environment. The Recovery Time Objectives (RTOs) vary, but are consistently the fastest of any solution in the industry. We often hear about entire data centers failing over in about 30 minutes.
Having a history of such unrivaled RPOs and RTOs, we think its important to be clear about service level expectations when recovering to AWS in ZVR 4.0.
The design, capability, features and the RPO of a few seconds in ZVR to AWS is exactly the same as what you find in an enterprise or managed cloud deployment. However, the service level for Recovery Time Objectives (RTO) is longer than what you find in an enterprise deployment of ZVR. We are seeing about an hour RTO when recovering VPGs to AWS.
For many workloads, that is a perfectly acceptable, so this addition gives decision makers even more options when designing solutions to meet SLAs. Some workloads are a great fit in AWS and don’t have the aggressive RTO requirements of just a few minutes.
Finding the right solution to meet SLAs is easier with Zerto because you can connect as many dissimilar environments as you need. Zerto has always been hardware agnostic, and with ZVR 4.0, we are becoming hypervisor agnostic. The addition of AWS adds even more options for protecting data. So if you require lower RTOs for some VMs, a private cloud deployment or working with a managed service provider is a better fit and you will get the lowest RTOs in the industry. For other workloads that have less aggressive RTOs, you can send them to AWS.
Stay tuned for more information about Zerto for Amazon AWS!