One of the major obstacles to the concept of cloud computing and “on-demand” is implementing the “on-demand” piece of the equation. Virtualization in theory allows organizations to shuffle virtual machine images of applications to and fro without the Big Hairy Mess that’s generally involved in physically migrating an application from location to another. Just the differences in hardware and thus potential conflicts between hardware drivers and the inevitable “lack of support” for some piece of critical hardware in the application can doom an application migration. Virtualization, of course, removes these concerns and moves the interoperability issues up to the hypervisor layer. That makes migration a much simpler process and, assuming all is well at that layer, mitigates many of the issues that had been present in the past with moving an application – such as ensuring all the right files and adapters and connections were with the application. It’s an excellent packaging scheme that migration as well as it does rapid provisioning.
The problem, of course, has been in the network. Virtual images aren’t small by any stretch of the imagination while Internet connectivity has always been more constrained. Organizations did not run out and increase the amount of bandwidth they had available upon embarking on their virtualization journey and even if they did, they still have little to no control over the quality of that connection. So while it was possible in theory to move these packages of applications around to-and-fro, it wasn’t always necessarily feasible.
Thus it is that solutions are appearing to address these problems to make it not only possible but feasible to perform migration of virtual images on-demand. NetApp Flexcache is just one such solution. Flexcache leverages data reduction and caching to ease the burden on the network of transferring such “big data”. Alone it is a powerful addition to vMotion, but it’s focused on storage, on the image, on data. It’s not necessarily addressing many of the core network issues that can cause a storage vMotion to fail.
That’s where we come in because F5 BIG-IP WOM (WAN Optimization Module) does address those core network issues and makes it possible to successfully complete a storage vMotion across the WAN. Application migration, on-demand.
Today’s F5 Friday is a guest post by Don MacVittie who, as you may know, keeps a close eye on storage and WAN optimization concerns and technologies in his blog. So without further adieu, I’ll let Don explain more about the combined F5 BIG-IP WOM and NetApp Flexcache solution for long distance transfer of virtual machines.
VMWare vMotion allows you to transfer VMs from one server to another, or even from one datacenter to another, provided the latency between the datacenters is small. It does this in a two-step process that first moves the image, and then moves the running “dynamic” portions of the VM. Moving the image is much more intensive than moving the dynamic bits, as the image is everything you have on disk for the VM, while the dynamic part is just the current state of the machine. Moving the image is referred to as “storage vMotion” in VMWare lingo.
NetApp Flexcache enhances the experience by handling the transfer of the image for you, making it possible to utilize Flexcache’s data reduction and cache refresh mechanisms to transfer the image for the vMotion system. While Flexcache alone is a powerful addition to vMotion, it does not address latency issues, and if the network is lossy, will suffer performance degradation as any application will.
F5 BIG-IP WAN Optimization Module (WOM) boosts the performance and reliability of your WAN connections, be they down the street or on a different continent. In the datacenter-to-datacenter scenario, utilizing iSessions, two BIG-IP WOM devices can drastically improve the performance of your WAN link.
Adding F5 BIG-IP WOM to the VMWare/Flexcache architecture provides you with latency mitigation techniques, loss prevention mechanisms, and more data reduction capability. as shown in this solution profile, a VMWare/Flexcache/WOM solution greatly increases the mobility of your VMs between datacenters. It also allows you to optimize all traffic flowing between the source and destination datacenters, not just the vMotion and Flexcache traffic.
While the solution involving Flexcache (diagrammed in the above-mentioned Solution Profile) is more complex, a generic depiction of F5 BIG-IP WOM’s ability to speed, secure, and stabilize data transfers looks like this:
|Related blogs & articles: |