On the surface, security questions surrounding virtual servers don't seem much different than those for the physical machines on which they run. In fact, starting a virtual security audit by keeping in mind what you've already learned in the physical world is an excellent approach. Security analysts say the same practices, principles and basic common sense apply for a group of virtual servers as for any physical server farm. But, IT managers also need to factor in some additional considerations, due to the unique characteristics of the virtual world.
One example: Software can be deployed so much more quickly using virtual machines that some steps in the typical provisioning process may have been eliminated, says Paul Love, director of information security at Standard Insurance in Portland, Oregon. That, in turn, requires IT departments to make sure the necessary controls and oversight are in place, with the truncated time frame in mind.
"With virtual machines, it's very helpful to pay attention to the actual configuration of the system," Love says. "You need to really have a stable build so that when you deploy a thousand versions of it, they all meet management's requirements for what controls should be in place."
When Love's team audits security for its virtual server environment, it doesn't introduce new steps so much as extend the ones it already has for physical servers, Love says. That includes looking at the interactions among systems and ensuring that the operating system on which the virtual machine runs is secure and encounters no "configuration drift."
"We have to work very closely with change management," Love says.
As background research for auditing and improving your virtual security, you may want to consult guidance for securing virtual server environments that's available from the Center for Internet Security, the Defense Information Systems Agency and virtual server leader VMware.
"They [IT leaders] need to read these guides and come up with a summary set of lock-down and hardening policies that are customized for their environments," says Nand Mulchandani, senior director of product management and marketing at. VMware. "If you just do that one thing, you will be vastly more secure and safe."
Virtual security tools can also help, but analysts warn clients to first consider the products they already use before buying new ones specifically designed for virtual servers. There are already 10 to 15 vendors offering VM-specific security tools, and that figure will probably rise to 30 by year's end, says Chris Christiansen, an analyst at IDC.
Consider this five-step checklist when securing a virtual server environment:
1. Conduct a full risk assessment to understand how resources have been separated and aggregated.
"Don't forget what you've learned about risk management and configuration," advises Pete Lindstrom, an analyst at Burton Group. "There's not a whole lot that has changed drastically, except it's sort of mind-expanding to consider the notion that a physical host has an entire network segment sitting inside it. That means you need to evaluate the configurations of the virtual machines themselves. Do that the same way you do any other configuration audit."
2. Validate the process for creating, deploying, managing and making changes to virtual machines.
This is particularly important now that steps such as procuring hardware, loading the operating system, testing and arranging for rack space are no longer required, saysJohn Pescatore, an analyst at Gartner.
"It's very important that virtual machines don't necessarily belong to one group within an organisation," adds Standard Insurance's Love. "From a security standpoint, it helps having a dialog with the people administering the systems and the network group to understand what's changing in the virtual environment."
Citing an example of a potential problem, he adds, "You could have a virtual machine come up and be taken offline before you run your scans."
3. Securely configure the virtualisation layer and keep patches up to date.
"Read the hardening guidelines as a starting point to develop a baseline and then audit against that to ensure the security of the virtualisation layer hasn't drifted," advises Neil MacDonald, an analyst at Gartner. Tools from vendors such as Configuresoft and Tripwire can help with configuration, MacDonald says.
4. Secure the internal virtual switch inside the virtual server. Weigh the need for additional controls such as a virtual firewall or virtualised intrusion protection system.
"You have to pay attention to how your VMs are communicating with each other and with the outside world, and that means through the virtual switch infrastructure on the physical box," said Lindstrom. "You have to pay attention to the configuration of those switches and the movement of traffic and the accessibility to that traffic, from VM to VM and from VM to external devices."
5. Exert tight controls on access to the service console and management tools
You want to thoughtfully control access to consoles and tools such as VMware's VMotion and Virtual Center, and their equivalents in other environments. Best practice calls for management tools to run on a separate network.
"This will make sure the virtual machine cannot snoop into the traffic that your management console is having with the servers to control them," says VMware's Mulchandani. "It's almost like tapping into someone else's phone."
"Most security problems in the virtual world will be introduced through misadministration, mismanagement or just plain old mistakes," says MacDonald. "The fact that we have to use different tools in the physical world than in the virtual world compounds that problem."
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.