This post was written by Zerto’s Sr. Marketing Architect, Shannon Snowden.
One common question we get from service providers, enterprises, and others relates to how Zerto’s Virtual Replication (ZVR) is about data replication in large environments; how is Zerto able to protect servers with very large storage requirements?
Servers with terabytes of data are very common in a ZVR deployment because we are protecting so many higher-level critical servers. Because we protect those servers so well, we also often get questions around how ZVR protects very large, even up into the petabytes of data.
It really comes down to ch ch ch..changes.
Changes happen on the VMs in virtual protection groups (VPGs). A ‘VPG’ is the name Zerto gives to affinity groupings of virtual machines, usually by application or other similar functionality, into a single entity for replication. A VPG allows for boot order customization, startup and shutdown scripts if necessary. You usually want to put VMs together into VPGs that have some association with each other.
Planning for the Rate of Change
Zerto provides planning calculators in our customer portal. One tool is the WAN bandwidth estimator tool. The WAN calculator tool gives an estimate of required bandwidth between the source and recovery sites based on your expected data change rate.
The change rate with ZVR is the important factor as opposed to the total amount of data being stored. You can have petabytes of data but if only a fraction of that is changing then it really is inconsequential to the actual daily replication bandwidth requirements once the initial sync has completed because, “some things ain’t ever gonna change”.
For example, if a server in a VPG that has 1TB of data and 5% of the data changes per day, that is 51.2GB of data that is actually changing. Therefore, 51.2GB over 24 hours is 621.4 KBps. Using the Zerto WAN sizing calculator, the VPG would use 4.86Mbps uncompressed for the WAN connection. If you use ZVR’s built-in WAN compression, an average we often see is the bandwidth usage will be about 2.5Mbps.
But you wouldn’t want to try to replicate terabytes of data across the wire, so ZVR uses a pre-seed feature to sync the initial data set.
Changes matter more than the total amount of data associated with a server when ZVR is doing the replication and that is one of the reasons ZVR is so efficient with bandwidth usage between sites.
…our next post will cover some details on the pre-seed functionality, stay tuned…