CONNECTICUT LIBRARIES
April 2003, Vol. 45, No. 4
The name of the town is Beatrice (Bee-AT-ris), Nebraska. Outside of that little heartland town, in the mid-1960's, existed a school named Pershing College. Its campus, comprised of four new buildings in the shape of flying saucers, sat among fields of corn and milo. I will not bore you with how I came to attend that college in its first year of operation. Most of the students were from "back east." A handful of us were from the New York City and Long Island area.One night, a guy sought me out and announced that he and I were "sort of related by marriage." Maybe I should have been amazed, but I was not. This sort of thing had happened to me before, though never in the middle of nowhere! My parents' families were large and, when I was younger, I was often introduced to strangers who, I was told, were vaguely related to me. I had learned about networks early in life.
Author Barabasi is a Notre Dame physics professor who has studied networks as phenomena outside the field of physics, as well as within it. His book attempts to explain networks, their laws of operation, and why they are a part of almost every field of study: computers and electronics, certainly, but also sociology, business, biology, etc.
Networks are always comprised of nodes and links. Nodes are units, points, or locations. Nodes are attached or related to each other through links. A single node may be linked to another node or to several, depending on the structure of the network. A third component of a network is a hub, a "super" node that is linked to many, many more nodes than a typical node. While not all networks have hubs, those that do are called scale-free networks, and such networks are remarkably common.
Consider computer networks. We recognize PC's, printers, and scanners as nodes on our LAN. We scatter those hubs about our architecture to link several nodes to each other and to the server. On the World Wide Web, home pages are our nodes. Some pages, however, are more equal than others! Google and Yahoo are linked to thousands upon thousands of pages, while your home page and mine may be linked to only a few other sites.
Networks are at work in social situations. Barabasi notes that close friends and relatives rarely refer job hunters to job leads. Rather, it is usually a friend of a friend or a distant relative who passes along the lead that materializes into a job. Barabasi explains that our closest friends and relatives know us and they know each other, so it is unlikely that they would know of a job opening that we too would not be aware of. The "hubs" among our friends and relatives hear about opportunities and pass them along.
When I was CLA's PEG Chair, shortly after the program first started, I had to peddle buttons to raise funds at the annual conference. I was uncomfortable about approaching people who did not know me and did not know PEG. A colleague suggested that I ask a mutual acquaintance to help me out. She reasoned that this gentleman had not only worked in Connecticut for decades, but had also taught at Southern for many years. He knew dozens of librarians and, she was confident, he would enlist some of them to sell buttons. When the Conference was over, the coffers were full and people were buzzing about PEG, thanks to the network!
Scale-free networks exhibit "topological robustness" or tolerance against failures. Because a network structure provides multiple links among its nodes, a failure in one or even several links does not cause the network to malfunction. This is why the Internet stays "up" even when many servers, routers, and other pieces of hardware may be down. However, scale-free networks are vulnerable to attack. Vulnerability is related to dependence on hubs. Target a hub and you destroy large numbers of links, links that the network has grown to rely upon to carry out its functions. Networks with no hubs are less vulnerable to attack, but far less efficient under normal conditions.
Scale-free networks appear in nature and in man-made networks. They are usually the result of continuing growth and "preferential attachment." This latter term describes the development of hubs. It refers to a tendency of nodes with the most links to acquire even more links at a faster rate than other, more typical nodes. Generally, older nodes have more links than newer nodes.
Barabasi sees networks everywhere. He convinces us that understanding networks is vital to understanding our world. If you do not believe him, ask anyone from a large family.
Return to Looking at Books Online