Service deployment models — from bare metal to virtualization, containerization, serverless, and cloud — directly shape how networks are designed and managed. Understanding these models is essential for building infrastructure that is secure, scalable, and reliable.
Service Deployment Models
Another thing that can affect our topologies is what deployment model we are looking at. There are several different ways we can deploy our services.
The main goal of all of this networking is to provide some sort of services. Maybe it is web services, email services, file services. There is a whole bunch of different services that we might be trying to get through all of this networking. But how we deliver this will really make a difference in how we set up our networks.
There are a lot of different ways we could set up these services. Initially, we would just set things up with bare metal. These right here are representing the different services - it is a serving plate, different services that we want to operate. This is installed on an operating system, which is installed on this hardware right here. But over time we have gotten to virtualization, containerization and serverless. What do these actually mean?
Initially, we would just install things on bare metal. I would go out and purchase a server, and I would set it up with the services I wanted. So maybe I set it up with directory services, DHCP, DNS, file services, web servers, a database server, and maybe some third party. Now all of the devices on this network can get services from this server.
We did run into an issue, though. By installing everything on one server, we caused problems. Here we have some sort of third-party application that maybe stopped us from being able to update the database, or the web server, or one of these other services, and now these services are old and the whole server is at risk.
So there are some advantages to splitting apart these different services onto different servers. Let us buy three servers now. On one server we are going to put some common services together that we find - directory services, DHCP, DNS and file services. Here we have a web server and a database server; maybe it is some sort of LAMP server, which is very common out there. Then we also have this third-party application that we install on another device. Now they are not interfering with each other, so we have this stability. We have created more security with this.
But some disadvantages to this are that it has been costly. We have had to buy a lot of different equipment here. It is not very scalable, because every time we roll out a new service we have to buy a new piece of hardware. It makes it harder to manage, and it just does not make us very agile - that is, we have to go through the whole purchasing process, which can take a while, just to roll out a new server.
Virtualization helped a ton, and even though it has been around for a long time, it became very popular, because what we could do is not buy additional pieces of hardware for this but still have separation between our services. We would install something like a hypervisor, and then different virtual machines, and install these services on those virtual machines. Now we created separation, giving us security, scalability, manageability - a whole host of benefits to using virtualization to roll out these services.
Other technologies came out as well: containerization and serverless. Containerization is very similar to virtualization with a hypervisor, except we are not duplicating the OS. With a hypervisor we duplicate the OS, making much more bloat, much more extra stuff that we are storing and doing. With containerization we are not duplicating the operating system, but we are still creating some isolation for those services. So it is kind of a hybrid between the full hypervisor that we are using and the bare metal that we would be using.
There is also serverless, which just allows us to really do away with all of this infrastructure side and just run our scripts. That was some additional technology that came out and helped propel us forward with creating a more dynamic way of delivering these services.
This all allowed us to say, well, why do we even need to have the server altogether? Maybe we can move these services up to the cloud. And that is exactly what has happened. Our bandwidth has gotten greater and greater, so it just makes sense that we do not host our own stuff, but we put it in the hands of professionals that do this on a day-to-day basis. Then we do not have to manage a lot of these, especially things like web services, databases and third-party applications. A lot of times it just makes sense to go hosted and have it up in the cloud.
So now we have options. We have options to put it here locally on a local server, or we can have it up in the cloud. By doing this, it does change our network requirements, because now this user, which used to have a direct connection to this server right here, needs to make sure that they have a stable connection to whatever service we are running up in the cloud. We need to make sure that this connection is solid. We are going to have more redundancy. We are going to make sure that there is more equipment, so if one piece of equipment fails then it will be taken over by another piece of equipment. So it is really going to affect how we design our networks.
Really, a lot of us are going towards hybrid systems. That means that we have some services up in the cloud, we have some local, and we have some that are up in the cloud and local that do some sort of syncing.
TechKnowSurge builds IT and cybersecurity professionals through hands-on, concept-first training built around real understanding — not memorization. Free interactive tools, structured programs, and 25+ years of real-world experience, all in one place.
Explore free tools and programs →