When virtual platforms became a common technology in data centers, many administrators were smitten with the ability to move virtual machines between physical servers, and rightly so. It was trivial to build a server at your desktop and then deploy it without mounting a new server or running new cables.
Companies like VMware, with its DRS technology, took virtual machine mobility a step further by moving virtual machines within the data center as needed based on resource demand. Today, rather than build individual N+1 servers to wait for failure or resource depletion, administrators can build out to their base level of need and let the virtual platform clone and scale up virtual machines as needed.
That same architecture can be used to augment geographically distributed data centers today by distributing virtual machines between locations. Admittedly, IT is still in the nascent phases of dynamically moving virtual machines between data centers based on resource demand, but virtual machines can be cloned between virtual data centers in different locations today. These dynamic virtual machines can be used to deliver applications and services to users in a more intelligent manner than the typical active/standby disaster recovery solution. In other words, dynamic mobility between locations may still be bleeding edge, but static virtual machine mobility between locations is available today and can offer new architectural models for virtualized applications.
A perfect use case for leveraging virtual machine mobility for intelligent application delivery across multiple data centers is geo-location: distributing a user, device, or service to the application that’s closest to them based on proximity. Geo-location offers multiple benefits to administrators and users, such as faster access to the applications because they are closer to the users, more sophisticated security and control over which users access which services and from where, the ability to create a follow-the-sun architecture for call and support centers…the list goes on and on. Geo-location brings the entire data center closer to the user, but it’s not really a virtualization technology; geo-location is more of a networking and infrastructure technology. Virtual machine mobility enhances geo-location by adding resource context to access decisions and availability.
Turning up multiple physical data centers across the globe has always been burdened by cost, time, and ROI. It’s expensive to bring up an entire data center with physical servers, cabling, headcount to deploy and manage, or the overhead of leasing space or turning to a managed provider. Adding geo-location at the network and application level typically required a large physical infrastructure to support distributing the user load between those locations. Deploying data centers across countries and continents was reserved for only the largest – or the most specialized – companies, drastically limiting the viability of using geo-location for proximity access. Virtualizing the data center and taking advantage of virtual machine mobility removes many of the barriers affecting geo-location.
If we look back at how virtual machine mobility helped spawn the massive growth in enterprise virtualization, we can apply that same model to a globally distributed network of data centers. Entire data centers no longer need to be physical, leased locations; now they can be virtualized in shared locations on existing physical hardware or in cloud locations. Administrators can still create virtual machines at their desks in San Jose, but now they can push those virtual machines to anywhere on the globe at will. One master virtual machine can be copied to the Bangkok virtual data center and then replicated on site, helping to mitigate WAN-related issues such as cost and latency. New services to take on increasing load can be provisioned within each data center as needed, recreating the scale up model in the now-virtualized local data center.
When geo-location is added to this model, dynamic provisioning and virtual machine mobility really takes off, and we start to see the underlying power of virtualizing all of our data center resources. Geo-location combines the best of both scale up and scale out worlds. With geo-location, decisions can be made not only on which data center to send a user based on location, but also on which data center can most efficiently scale up to handle the load. It can also make decisions on scaling out in a particular region with new virtualized data centers as need fluctuates, adding new flexibility to the typical follow-the-sun model. Geo-location allows context to be applied to application delivery, and virtual machine mobility allows that context to scale up and scale out dynamically as needed anywhere in the world. The enterprise can push applications as close to the user as possible, when needed, where needed, and as needed, without the huge up-front capital investment required to build a new data center on every continent.
TMCnet publishes expert commentary on various telecommunications, IT, call center, CRM and other technology-related topics. Are you an expert in one of these fields, and interested in having your perspective published on a site that gets several million unique visitors each month? Get in touch.
Edited by Jennifer Russell