8.3 Looking ahead
The last eight sessions have been based on home networks, or the networks you would find in a small business. In large organisations or enterprises, however, the equipment used is more sophisticated, and networks are more complex because of the use of subnetworks. The next eight sessions focus on ‘enterprise’ networks and processes. For example, setting up a router in an enterprise network is very different from setting up a gateway in a home network. The following video looks briefly at some of the significant differences between home and enterprise networking.
Throughout these eight sessions we been looking at networking, IP addressing and related ideas in the context of home networks. Actually, what we’ve been showing you wouldn’t just be found in homes. It would also be found in small businesses, such as shops, cafés and restaurants, solicitors’ offices, estate agents, and so on.
In large businesses and organisations, the same basic principles of IP addressing, subnet masks, DHCP, network address translation, and so on, apply. But the way they are implemented, for example the equipment used and the network structures, are more complicated. In fact, organisations like these to employ full-time network managers to keep things working.
Networks at this level areis often described as ‘enterprise’ networks, and the equipment used is described as enterprise routers, enterprise switches, and so on. Setting up enterprise equipment and networks is more involved than it is for home equipment and networks.
The next eight sessions relate mainly to the enterprise context. In this part I want to give a quick overview of the way enterprise networking differs from home networking.
As we’ve already mentioned, the home gateway, which is often informally called a router, combines the functions of a router, a switch, a DHCP server, and a Wi-Fi access point.
In the enterprise context, these are usually separate pieces of equipment, and often far from each other.
In a home gateway, the router has just two interfaces. Setting up these interfaces was mainly a matter of entering appropriate numbers in a graphical interface.
An enterprise router might have two or more interfaces. Giving an interface an IP number on one of these routers is usually done through a command-line interface, and involves several steps. It’s very different from the way you do it on a home gateway. Just to give you a flavour of what’s involved, I’ll set up an IP address on one of the router interfaces here, using the command-line interface.
I go to the command prompt of the router. I type enable. The prompt changes to a hash sign.
The next stage isn’t actually necessary, but it’s useful: show ip interface brief. It shows that there are two Ethernet interfaces, without IP addresses, and both are ‘down’, that is, not working.
So now I type a new instruction, configure terminal, and the prompt changes again.
Now I choose which interface to set up. This is the gigabit Ethernet 0 slash 0 interface.
The prompt changes again, and I key in the IP address and the subnet mask.
And then this curious instruction, no shutdown, to switch it on, or ‘up’.
Now it’s useful to go back to the show ip interface briefcommand, just to check everything looks right. And there we see that the 0 slash 0 interface has the right IP address and is ‘up’, that is working. So all that’s just for one interface.
Another distinctive feature of enterprise networks is the extensive use of subnetworking to create manageable groups of users.
For example, in a large organisation, you could have subnetworks for these groups of staff, and within each of these groups there might be further subnetworks. Typically there might be differing security policies for these subnetworks. For example, some subnetworks might only be accessible to certain categories of staff.
There might be further subnetworks for functions like printing, and subnetworks for servers providing public-facing services such as web pages, online purchasing, enrolment, and so on.
Enterprise networks also often span multiple sites, often widely separated, but interlinked so that, for the staff, there’s no difference between communicating with someone on the same site, or someone on a different site, possibly in another country.
So the world of enterprise networking is quite a step up from what we’ve been looking at so far.
Activity 3 Test yourself
1. In the video, an enterprise router is described as having two or more interfaces. This refers to the number of network interfaces. Why might such a router have more than two network interfaces?
To transfer data between more than two networks. A router’s function is to act as a gateway between networks, so that data packets can be transferred between networks. A router with four network interfaces can receive a data packet on one interface and forward it to one of the three networks connected to its three other interfaces.
2. The activity shows the interface of an enterprise router being set up manually via the command-line interface. Enterprise networks use DHCP, so why is manual configuring of some interfaces necessary?
On the public internet, network addresses need to remain constant so that packets can be directed to the right destination network. This means that router interfaces need to remain fixed.
3. Towards the end of the video, it is pointed out that some enterprises have an international spread, and yet ideally users’ experience should not be noticeably different when communicating with remote colleagues compared with local colleagues. What particular difficulties could be expected with such long-range communication (compared with local communication)?
One problem is increased latency for long-range traffic, which almost certainly has to pass through more routers than local traffic. Another potential problem is security, as data could be expected to be vulnerable on the public internet. (In fact, so-called ‘tunnels’ are used in practice to mitigate both problems. Data tunnels are virtual private lines, which nevertheless operate over the public internet, and data traffic can traverse them with less latency than ordinary IP traffic. Traffic in data tunnels is also usually encrypted.)