Author: Mike Talon
What happens if you have lots of bandwidth, flexible Recovery Point Objectives and you’re virtualized both in production and DR? Well, you can take advantage of Double-Take Virtual Host Edition for a start.
Double-Take Availability can easily run on virtual guest devices, both Windows and Linux. You can protect those guests either to other guests or (for Windows only) to a Virtual Recovery Assistant helper VM. That’s the solution with the most flexibility in the product line, but some customers don’t need that level of flexibility, and would rather replicate at the ESX or Hyper-V host level instead.
Double-Take Virtual Host Edition is the perfect solution for that host-to-host replication. Formerly known as Double-Take for Virtual Infrastructure, the product got a rename when the 5.0 version was released to make it sound more like what it does – replicate VMs via virtual hosts.
It supports ESX 4.x and Hyper-V RTM and R2, including the Hyper-V Role on Server 2008 and 2008 R2, and is designed quite differently on each platform:
On ESX, Double-Take Virtual Host Edition leverages native VMware snapshot technology to take periodic snapshots that hold up to 15 minutes or 32 MB of data – whichever comes first. By closing one snapshot and immediately opening another, we can provide continual protection for any VMs on that ESX host, and transmit the now closed snapshot to a target ESX device. Once on the target, the snapshot is committed to a duplicate copy of the VMDK files for those protected VMs, and the original is then removed from the source ESX host. When the next snapshot window closes, the process repeats. Since we also re-create the VM definition files onto the target host, if the host or a VM fails, we can bring up those VMs on the target device.
This is great for those shops that have bandwidth to spare, and would prefer a more host-centric approach to High Availability. The ESX architecture does enforce some restrictions, however. This is not strictly real-time replication, though it can be quite close with the rightinfrastructure. You will also need to be sure the hosts are sized properly to allow the system the overhead required to do the periodic snapshots. Finally, you’ll want to make sure you’re using a stretched-LAN or virtual LAN in order to allow for VMs to re-connect on the DR site. If, for any reason, these restrictions don’t work for you, no problem! Just use Double-Take Availability for Windows or Linux within the guests; and all of these restrictions disappear.
On Hyper-V, the Double-Take Virtual Host Edition system works much more like the Double-Take Availability product line. Since the Hyper-V host (both Hyper-V Serverand Role) is a Windows Server 2008 or 2008 R2 device, we can provide true byte-level real-time replication over much lower bandwidth than we can provide at the host level on ESX. This means you are essentially performing Application-Level Failover (similar to DTAAM forExchange or SQL), where the Hyper-V VM is the “application” being failed over. We replicate all VHD’s and the virtual machine definitions for those VMs you want to protect, and can failthem over just like any other application.
Double-Take Virtual Host Edition gives you even more options for protection of a virtual infrastructure on VMware ESX or Microsoft Hyper-V. Knowing that you have the ability to do guest level (Double-Take Availability) or host level (Double-Take Virtual Host Edition), means that you can use one suite of products from Vision Solutions, Inc. to protect anything in your physical and virtual environments.
