Creating a new network can be daunting, but assessing the hardware needn’t be. Regardless of the size of your organization — from a few seats and no centralized server to anything larger — certain fundamental rules apply across the spectrum of possible network configurations.
The examples below focus solely on hardware implementations, giving no special attention to budget or any one vendor over another. Budget and vendor choice are separate issues that should be addressed before buying a single piece of hardware. Within the scope of this article, we’re talking about building a network from the ground up.
We also assume that you have scoped the correct number of devices to network — including workstations, laptops, and other hardware — and they are ready to go. Again, we’re focusing on the network implementation itself.
Before you begin: Some basic tools
Before getting to the main hardware itself, you'll need some basic tools on hand. You'll want network testing equipment to “tone” lines. The toning process means sending a single signal down a network cable — even if your devices are wireless, at some point in the process you’ll have to install cabling — to find out where it goes.
If pre-fab network cables are available, then those plus cable ties will be the essential tools.
If you are building your own network cabling, you'll also need wire cutters, Category 6 network cable, crimping tools, cable ties ... All these are essential to have on hand to set up the network hardware properly.
Setting up switches and routers
Once you’ve determined where your network closet is going to be, you should outfit it with one or more switches. These are “traffic directors” for your network. Unless your organization is very small, and you don't expect it to grow, switches with multiple available ports are best!
From the location of the switch(es), run cables to any routers you need to manage network traffic. Even if you plan to connect devices wirelessly to the router(s), you’ll still have to get data to them from the switches. And a bit of advice: Label your cables! We’ve all seen pictures of racks and patch panels that are a nightmare of unmarked, tangled cables. Save yourself a great deal of heartache, and right off the bat start by using clearly labeled cables, marking what they go to (and from) and tie them off neatly once they’re in place!
Connecting the modem
Depending on your connectivity to the outside world (if your organization has or requires it), at this juncture you should connect your modem(s). Ideally, the modem or modems should be located in an easily accessible location, preferably one that is close to both network and IT support. That way, if you have to service or walk through troubleshooting from your bandwidth provider, they'll be close at hand.
Internal network management
Once basic connectivity has been established, and core infrastructure has been laid in (again, either wireless, cabled, or a combination of both), you need to address internal network management.
Here you have some interesting options, depending on the size and scope of your organization. First, establish a domain controller. The domain controller authenticates users and devices on the network and permits them access to network resources and connectivity to other devices on the network. Next, if necessary, set up a DNS server be to manage Domain Name Services. Finally, establish a DHCP server to manage internal IP distribution.
All of these devices can be created on one or more stand-alone server-class systems, either a rack-mounted or stand-alone server. While not necessary, a single computing device can be devoted to each task. Alternately, each of these duties can be virtualized (that is, run on a virtual machine on a single host). For small organizations that only require access control via a domain controller, and DHCP management, many routers do offer those features within their own firmware. This means the entire network can be managed from a single, simple “box.”
While it simplifies the system, running everything from a single box also creates a single point of failure, and a single point of weakness. If it malfunctions or quits working altogether, all network traffic grinds to a halt. Worse, it can be quietly compromised, either through lax security (login: admin, p/w: admin are popular defaults — please please please change them when you’re configuring your routers!) or by brute force attacks, human engineering, and so on.
Establishing a robust password policy
This brings up a very critical stage of network -building: Once constructed, a robust password policy must be established! Whether your organization makes tie clips or jet engines, it is subject to outside interference. Even if malefactors aren’t interested in what you’re doing, subverting systems on your network for botnet attacks is still a danger.
Choosing which option(s) work best for your organization is, ultimately, up to you. There is no one be-all, end-all network startup configuration that is perfect for every organization, but the framework, the backbone of it all is the same whether it is a three-employee office that will never grow or 30 thousand employees in offices across the country or the globe.
