There is no doubt that organizations are embracing server consolidation to improve efficiency and save money. But although the footprint of the primary data center may be shrinking, getting essential services and applications to branches and partners usually means deploying a remote server. This tip helps IT administrators understand how to manage and maintain those remote servers quickly and efficiently.
Driving factors behind remote servers
The need for remote servers is driven by the need for business software performance and practicality. It is, at least theoretically, possible for a server to provide services for a global user base, but achieving an acceptable user experience at an economical price point can be a considerable challenge.
It’s important to note that remote server management doesn’t automatically imply that an organization deploys distributed servers. An organization may successfully consolidate its servers into a single location, and chances are those servers will be managed remotely on a routine basis.
For example, a complete “lights-out” data center may exist in one place and may be managed entirely from a remote or even a mobile location. This is exactly the case with Ian Parker, senior Web services administrator for Thomson-Reuters, a business intelligence company based in Ann Arbor, Mich. The move to shift distributed servers to a central location and central management started long before virtualization. “For at least five years now, I haven’t touched one of my servers unless I take a road trip out there,” he said.
Remote servers can be managed using a variety of technologies, each dictated by the unique needs of a specific organization. For simple management tasks, such as software maintenance upgrades and status checks, a tool like Windows Remote Desktop is often adequate. But software tools work only as long as the server being managed is operating normally. When an application hangs or the server experiences a hardware problem, software tools can be ineffective.
Remote management includes hardware-based offerings
Hardware-based tools, such as HP Integrated Lights-Out (iLO) and Dell Remote Access Controller (DRAC) cards, are a bit more foolproof, allowing a remote administrator to monitor, configure and control more of the server’s hardware elements over a network as if the administrator were actually sitting at the local console. Many blade systems also incorporate remote management features that provide remote administrators with Web-based access to diagnostics and almost total control over the physical box.
KVM-over-IP technology is another hardware-based means of viewing the remote server’s console and controlling the keyboard and mouse input to that server from an administrator’s location over an Ethernet network. An increasing number of UPS systems also include remote management capabilities so that an administrator can cycle power to the server if necessary.
Server virtualization has little direct impact on remote server management. “I’m doing that the same, whether that server is 50 feet or 500 miles from me, whether it’s a physical box or a virtualized box,” Parker said.
But the addition of virtualization does provide more control and options than traditional non-virtualized boxes. One example is live migration of virtual machines (VMs). A remote administrator can move VMs between remote physical servers for maintenance or to minimize the impact of a remote server reboot. Similarly, new VMs can be provisioned remotely with the same speed and ease that is experienced locally. Ultimately, virtualization adds important flexibility to remote management, but the underlying goals of remote server management remain unchanged.
This was first published in August 2011