Network Topologies in E-Commerce

Updated About content Print Article Share Article
views updated


When setting up the computer systems for your e-commerce enterprise, you need to figure out how your computers will be configured in relation to each other. Your computer systems need to communicate with each other in some fashion based on how many people are working together and what kinds of computer files and resources they share. The way they are tied together in your computer network is known as a network topology. There are three primary topologies common to most computer networks, and you can either adopt one of them or combine different elements of each.


The oldest and most common topology is the star network. In this configuration, a central computer, often a powerful one, resides at the figurative center of your network, and connects individually to a number of smaller workstations, thereby centralizing the network communications between the individual computers. The star network is the topology of choice for mainframe-based networks, where the mainframe links hierarchically to a series of microcomputers point-to-point and acts as the host computer for them. In this case, the mainframes perform the bulk of the work, while the workstations generally just send and receive data to and from the host.


Ring network topologies are similar to star networks except for one crucial difference: there is no central mainframe or host computer to which the workstations are connected. Instead, the workstations all connect to each other as if in a closed loop. More expensive than the alternatives, ring networks are probably the least common network topology. However, one advantage of the ring architecture is that, within smaller local area networks (LANs), data flows are much more fluid and change directions with greater regularity. A ring architecture allows for connections to be quickly and easily rearranged to account for the more dynamic data-traffic demands. This flexibility in the configuration of data flows also allows for greater adaptability, so that as you begin to combine network topologies, you can pick and choose the most effective entry points and hubs for connecting the ring with the outside network.


Bus networks take their name from urban bus systems, and function in much the same way. In this topology, workstation microcomputers are connected to each other via a single high-speed cable that runs through each computer station. The computers thus constitute "stops" along this cable route, and along this route data stops at each computer to drop off and pick up data. Thus, like ring networks, the bus network topology features no central host computer.


Depending on the size and complexity of your company, you can also combine a few different topologies, creating a larger topology out of two or more smaller configurations. For instance, you can combine several different star networks via one bus configuration, where the bus runs through the individual star networks' hosts to drop off and pick up data. Thus, while the star networks feature individual workstations connected point-to-point with their particular hubs, the bus cable runs through the successive hubs to combine the different networks. This method is known as "daisy chaining" the topologies together. Similarly, star networks can be combined into a large ring network via a high-speed fiber-optic backbone. In this method, the hubs of the individual star networks can act as, or plug into, a hub in a circular chain.


The bottom line is that your network architecture needs to be strong and resilient enough to ensure business continuity, avoiding outages and hang-ups. Customers in the fast-paced e-commerce market are notoriously impatient with downtime, perhaps seeing it as a sign of shoddy IT practices, and so your network topology should be designed not only for efficient data sharing, but for durability, as well. Thus, effective management of your internal network architecture is a crucial component of your customer service operations, both in keeping the network running and in ensuring the greatest possible speed and efficiency of data flows through your organization and between you and your customers.

This also impacts your method of data backup. While you many wish for the convenience of storing information in a central mainframe computer, the loss of that information due to a mainframe crash or severe virus of some kind could be devastating. To ward off such possibilities, many firms opt to back up their data using network-based methods such as online disc mirroring, where information is sent via your wide area network (WAN) to various remote data centers, protecting your information by dispersing it.

At the most basic level, a physical map of your network is useful to analyze and manage the data flows through your company's network infrastructure. These maps are what give the star, bus, and ring networks their names, as the layouts rendered on a map will visually resemble those structures. You need to be able to clearly visualize the exchange patterns in your organization, and zero in on trouble spots to figure out the most effective way to overcome data-transmission bottlenecks and other information-flow inefficiencies.


Bament, Sally. "Rethinking Telecom Architectures." Business Communications Review, December 2001, 16.

Morrissey, Peter. "The Survivor's Guide to 2002: Infrastructure." Network Computing, December 17, 2001, 69.

Roy, Gerry. "Network Management Leads to Customer Satisfaction and Big Bucks." Computer Technology Review, December 2001, 26.

Rusin, David G. "Choose Your Routes Carefully." Telecommunications Americas, October 2001, 52.