Don't Forget The Network Connectivity Devices

Written by Julianus Yu

We’ve looked at the products you can use to bring your communication endpoints to a central location. But is there any communication taking place over your infrastructure? What you need now is a way to tie everything together, you need to establish seamless communication across your internetwork.

Repeaters
Nowadays, the terms repeater and hub are used synonymously, but they are actually not the same. Prior to the days of twisted-pair networking, network backbones carried data across coaxial cable, similar to what is used for cable television. Everyone would be connected to the same coaxial backbone. Unfortunately, when it comes to electrical current flowing through a solid medium, you
have to contend with the laws of physics. A finite distance exists in which electrical signals can travel across a wire before they become too distorted. Repeaters were used with coaxial cable to overcome this challenge.

Theoretically, repeaters could be used to extend cables infinitely, but due to the underlying limitations of communication architectures like Ethernet’s collision domains, repeaters were originally used to tie together a maximum of five coaxial-cable segments.

Hubs
Because repetition of signals is a function of repeating hubs, hub and repeater are used interchangeably when referring to twisted-pair networking. The semantic distinction between the two terms is that a repeater joins two backbone coaxial cables, whereas a hub joins two or more twisted-pair cables. In twisted-pair networking, each network device is connected to an individual network cable. In coaxial networking, all network devices are connected to the same coaxial backbone.

Hubs typically provide from 8 to 24 twisted-pair connections, depending on the manufacturer and model of the hub (although some hubs support several dozen ports). As a media-access architecture, Ethernet is built on carrier-sensing and collision-detection mechanisms (CSMA/CD). Prior to transmitting a signal, an Ethernet host listens to the wire to determine if any other hosts are transmitting. If the wire is clear, the host transmits.

Bridges
When we use the terms bridge and bridging, we are generally describing functionality provided by modern switches. Just like a repeater, a bridge is a network device used to connect two network segments. The main difference between them is that bridges operate at the link layer of the OSI reference model and can therefore provide translation services required to connect dissimilar media access architectures such as Ethernet and Token Ring. Therefore, bridging is an important internetworking technology.

Compared to modern routers, bridges are not complicated devices; they consist of network interface cards and the software required to forward packets from one interface to another. Bridging is one technique that can solve the shared-bandwidth problem that exists with hubs.

Switches
A switch is the next rung up the evolutionary ladder from bridges. In modern star-topology networking, when you need bridging functionality you often buy a switch. But bridging is not the only benefit of switch implementation. Switches also provide the benefit of micro-LAN segmentation, which means that every node connected to a switched port receives its own dedicated bandwidth. And with switching, you can further segment the network into virtual LANs.

Micro-LAN segmentation is the key benefit of switches, and most organizations have either completely phased out hubs or are in the process of doing so to accommodate the throughput requirements for multimedia applications. Although switches are becoming more affordable, their price may still prevent organizations from migrating to completely switched infrastructures. At a minimum, however, servers and workgroups should be linked through switched ports.

Routers
Routers are packet-forwarding devices just like switches and bridges; however, routers allow transmission of data between network segments. Unlike switches, which forward packets based on physical node addresses, routers operate at the network layer of the OSI reference model, forwarding packets based on a network ID.

Routers enabled with the TCP/IP protocol and all networking devices configured to use TCP/IP make some sort of routing decision. All decisions occur within the IP-protocol framework.