so we've talked about some ideas of virtual ization. You know, we've talked about virtual machines, virtual desktops, talk about virtual lands and cloud access of resource is which really is kind of virtual ization as well. So any time we're thinking about any virtual ization solution,
we always want to make sure our very first concern is
cost benefit analysis. Does it provide us the degree of security based on our requirements and the value of the data? We have to do a lot of research here to make sure that we continue to meet our requirements. Certainly if we're governed by legal or regulatory compliance issues,
we've got a look at physical access
again. When we're sending this out to the cloud, perhaps we may not have as much control, but we still want to make sure physical security is present when we talk about virtual lands within our organization. When we talk about virtual servers within our organization, we still run this virtual ization on physical machines,
one of the tenets of security
availability. All the security in the world doesn't matter if I can walk into your server room and walk right out with that server and under my arm. All those protective mechanisms you've put in place have have done nothing's for is the availability of data. So we always want to make sure that physical access
is a consideration we make.
Don't forget. We've talked about layer defense, technical, physical and administrative. We always need to consider those elements
now, um, for virtual machines, one of the things we have to think about. We've talked about how that allows us to have thin clients, and that's very true. But the system that's hosting the virtual environment a lot of times we just call that the host system or the host OS. That has to be a very robust system
because it's running lots of many virtual machines on there.
So we have to make sure that we monitor the resource usage. This is a system that needs a lot of processing capability, needs a lot of ram, needs a lot of hard drive space, so whereas we're saving money on the clients and that generally outweighs the money we have to spend. But we've got to spend some money on that back end server.
We need to have set procedures in place, you know, change control, something that will talk about later and configuration management. You know, in the world of security, nothing should happen on the fly. We should be able to have a set of policies and procedures that we follow
to virtual eyes software to Virtual Eyes Service's.
And we should follow those guidelines. We have baseline security configurations for the host operating systems. We should also have baseline security configurations for the virtual machines as well.
Things like removing a necessary service is making sure that our systems air patched and run the latest. The latest version of the operating system, as faras is it's appropriate. Rename administrative accounts, change default settings. Those good, solid principles of security,
um, ensure any physical network switch. Ports connected to virtual switches are configured as static trunk links. Now, when we talked about switches, I really didn't talk about trunk ing or villains going across trunks. But the idea is
like most things in networking configuration is either static or automatic, sometimes called dynamic, and with static configuration, I have the greatest control, but I also have the most overhead
well, with dynamic configuration. You flip a switch, so to speak, and it works. At least that's the idea behind it. But any time you turn configuration over to an automated source, there's the potential for counterfeiting. There's a potential for spoofing.
There's the potential for any sort of rogue device to step in.
So when we talk about trunk ing with switches, we talk about, for instance, of villain being able to dynamically realize when additional switch has been added or an additional pork has been added to its feelin, and that's very beneficial from administrative state. But this idea static configuration
it doesn't matter whether you're talking about
trunk switches on how it deals with dynamic trucking protocol. It doesn't matter whether you're talking about th e P. Doesn't matter how you're talking about any automated service. You will always get greater security with static configuration.
So what makes the decision between automatic and static cost benefit analysis every time?
All right, make sure direct file sharing between host and guest operating systems is disabled, very important with virtual ization, because
there you know, a lot of times you install VM tools or whatever the little packages called for the particular piece of virtual ization software, but it allows sharing across the host and the VM, so it makes it very easy for me to download an application on the host system. Just copy and drag it to the VM where
but again, that I d expands beyond just its original use of making it easier to administer.
if I have a guest VM that connects US files on the host, you know what's to keep that from accessing other V, EMS or other sensitive that on the host we want. The whole purpose of virtual ization is isolation. We want to truly isolate each individual. Virtual machine
backup and recovery process is for your V EMS Absolutely, absolutely, absolutely
being able to restore of'em from a previous snapshot or image or whatever your technique is for backing these up. Because again, it's just another system. And don't forget you're putting many eggs in one basket because now I'm running. 10 different service is on a single physical box,
so I want to make sure if there's damage to that box.
If a virus infects the box, it's not unheard of for a virus to move from v MTV MTV M so I want to be very, very careful and make sure I have reliable backups
and recovery process. How do I test backups to make sure that they work to make sure that they're affected? There is only one way to test the backup, and that's to do a restore. And unless you are able to successfully restore the data from back up, you should have no faith in that backup whatsoever.
And unfortunately, sometimes companies figure out their backups weren't working
when they need to do a restoration.
Ah, one thing that I skipped here. Separate roles for system Add men's and network ad men's. That's a very important idea in the realm of separation of duties that shouldn't just be for virtual machines. That should be for your network Environment Network ad, Men's Do Security Add men's monitor.
You would never wanna have that be the same function,
right? Because if you think about all the things the network admin can do
pretty much anything. You know, if you're in a UNIX environment, you think about what the root conduce. The route is all powerful in Windows system at men all powerful,
so you would never want to hand over that degree of power and authority to a single entity without making sure that you have someone monitor what's going on there. That's the role of a security Admin is to monitor with the network ad men's do separate villains for host to guests and guest to guest communications
villains. Increased security.
So isolating out roles from network to network would add a degree of security
change management. I've mentioned that I'll continue to mansion mentioned change management configuration management.
Sadly, we as network people, computer people, I t people, security people,
we don't like to document,
and by we I don't mean us. Of course, you and I document because that's the right thing to do. But the rest of those people out there don't like to document. And so what that means is at any point in time that I might need to come in and reconfigure or recover a system. I need to have a reliable source to know with the configuration settings
off that system more
and I need to make sure that if changes are made to this system, those changes following orderly process that includes documentation that's not unique to virtual machines, that's for networking. That's for computers. That's for information. Security is a whole. We should always have a change management configuration management process.
Stay aware of risks.
Maintain consistency with your risk classifications. You know that that's not specific to virtual machines or virtual ization. Either we classify or evaluate our assets based on potential for loss and harm.
We want to maintain the grouping. It doesn't make sense. Tohave ah, high value in low value on the same system. So when we're thinking about how we're gonna group service is on a system. We would classify each system tau host something based on risk will have a higher set of security requirements
for upper level or higher risk or
higher value assets than we would for lower value.
Enable audit logs, yes, but also review audit logs. And don't forget audit reduction tools. And we talked about system event information managers where we have a means of taking this wealth of information because, like we talked about the target, they had 10
1000 entries in their law.
No one can make sense of all that information, so we need to utilize the tools that are available better out there in order to filter through and sift through what's actually meaningful.
enforce additional security control. Stir the use of purpose built firewalls and intrusion prevention and detection systems. What that means is continue the ideas of layered defense. You've got a server with high value data. Implement layers of security, not the least of which would be the limit physical access,
technical access By requiring at someone accessing that data to go through a firewall and make sure that they're legitimately allowed to do so. Make sure that that back in system might not be the subject of a denial of service attack. So basically a lot of thes rules and requirements apply to just
good foundational network security ideas.
And I think you'll see that throughout the exam. And one of the things I would tell you for the exam is when you get questions and you're not sure what the answer is, step back and just see if there isn't a universal security principle that you would apply to a given situation.
You know, they might say, What's the best thing you can do to harden a UNIX buying server?
Maybe you've been on a UNIX buying server. Maybe you have it doesn't matter. The best thing you do to harden a UNIX bind server remove unnecessary service is applied. The latest hot fixes and security patches
change administrative accounts. You know those sorts of things, so always try to step back and see if there isn't a foundational principle that you can apply separation of duties, auditing and reviewing the vault. It logs all those ideas that we just continue to talk about. So these air some ideas for securing with virtual ization.
I do believe virtual ization, something you'll see on the exam. I know you will.
I see a good handful of questions there. So, um, this is just a good review of some of the security configurations.