As IT environments evolve, administrators must adapt traditional practices to fit new technologies. With the physical-to-virtual model shifting to virtual-to-cloud, new questions about benefits and limitations arise.
Early limitations of virtualization had forced IT administrators to consider which physical servers made good virtual candidates. As
V2C faces similar issues as P2V; not all virtual machines (VMs) are suited to the cloud. Cloud admins should consider these four aspects before deploying any VM in a cloud environment.
More cloud management resources
1. Cloud-to-LAN networking.
One of cloud's biggest value propositions is its unlimited resource supply. Constraints on processing power, memory and storage can limit business agility; moving IT resources to the cloud can eliminate those strains.
The cloud model fails when network capacity between cloud provider and local LAN can't meet demands. A Microsoft Exchange server in the cloud might show impressive database performance but poor overall response time if the network between it and Outlook clients is insufficient.
The solution is proper monitoring. IT hasn't had to measure these metrics in the past, so they may be unsure how much network capacity a particular server requires. Successfully moving VMs to the cloud requires admins to carefully monitor network usage via third-party cloud and network management tools.
2. Data transfer costs.
Cloud computing’s unlimited capacity replaces the traditional connection between hardware and costs with a price structure based on consumption. Processing power, memory, storage use and networking are billed back to associated business units based on how much they use.
The consumption-based, pay-as-you-go cloud model can offer huge cost savings for IT environments that effectively manage resource use. Those that don't will quickly find the “hidden” costs that arise when resources aren't properly constrained.
Thankfully, predicting these metrics isn't difficult. Pay careful attention to how many resources your VMs use on an existing hypervisor first. In all V2C processes, the server begins as a VM, and all virtual platforms expose metrics relevant to the cloud-enabled VM's consumption calculation. Understanding those metrics will help you calculate costs and prevent an excessive monthly bill once those VMs move to the cloud. SolarWinds offers a free online calculator that helps you determine the cost to move VMs to Amazon EC2, Windows Azure and Rackspace cloud servers.
3. Cloud monitoring and verification.
Any cloud provider worth using delivers metrics on consumption and resource use for every hosted virtual machine, yet the rules of due diligence mandate an additional level of independent verification.
Prior to considering any major move to a cloud provider, plan for how you'll monitor VMs. That
extra effort ahead of time is important in ensuring your provider can deliver on your
4. Cloud migration tools.
Even with careful planning, the task of actually migrating a locally hosted VM to the cloud can be difficult. While P2V tools have become a commodity, few virtual-to-cloud migration tools exist.
Existing tools for migrating applications to the cloud from companies such as AppZero and Citrix’s NetScaler can be platform-specific and one-directional; some tools help move virtual machines to a cloud platform, but can’t migrate systems back to an on-premises data center. Pay attention to your tool choice and ensure you have a back-out strategy if you are forced to move machines or applications from the cloud back into the data center.
All things considered, the shift from physical to virtual to cloud appears to be an organic one. Few in virtualization's early days suggested every server might make a good VM. The same holds true with cloud hosting VMs. But technologies change, so we may eventually remember the cloud candidacy days no differently than we remember the virtual candidacy days.
Greg Shields, Microsoft MVP, is a partner at Concentrated Technology. Get more of Greg's Jack-of-all-trades tips and tricks at www.ConcentratedTech.com.
This was first published in April 2012