Steven C. Horii, MD
The essence of the DICOM standard is that it prescribes a uniform, well-understood set of rules for the communication of digital images. For the purposes of this article, communication is defined as the interchange of information. This sounds simple enough; indeed, it is something we do every day. However, the reason we are able to communicate successfully is that we follow a well-established set of rules that we have, for the most part, mastered from childhood.
Electronic communication is commonly thought of as being divisible into a set of layers with each layer performing a defined set of functions (2,3). This model of communication as a set of layers is part of an international standard for communication called the International Standards Organization Open Systems Interconnection (ISO-OSI) Reference Model (Fig 1). The model may be understood from an analogy to the structure of a manufacturing company. At the highest layer of electronic communication, the system interfaces to the user's application (eg, a computer terminal or a personal computer running a particular program that accesses data over a network). This layer corresponds to the planners and decision makers in the company who determine what product will be manufactured and shipped. At the lowest layer of electronic communication is the physical medium (eg, the "wire") over which the information will be sent and received. This layer corresponds to the trucks used by the shipping department. The designation of these layers as "higher" and "lower" does not imply levels of importance; rather, it is used because the layers are viewed as a stack with the physical layer on the bottom and the application layer on top.
In between the topmost layer and the physical layer are other layers that deal with matters such as what character set will be used to represent information, how to establish the rules for making connections over the physical medium, and how to handle any errors that might occur in the communication process. These intervening layers can be thought of as different departments in our hypothetical company, each performing specific tasks (eg, parts selection, manufacturing, quality control, planning shipping routes). In both the electronic communication model and the company, each layer (or department) accepts input from the layer above (preceding department), performs a well-defined set of functions, and provides output to the layer below (succeeding department).
The exchanges between layers or departments can go both ways, because most communication (as in the exchange of goods between companies) is bidirectional. In communications terminology, the movement of information between layers uses services that are provided by the layer. However, communication means the exchange of information; therefore, there must be a corresponding set of layers in another device to which the physical medium is connected.
Our hypothetical company manufactures and ships its goods to another company that incorporates them into its products. This second company does not ship products back to the first company (unless they are the wrong product or are defective) but sends back payment for the goods it has received. Expediting departments in the two companies may handle scheduling of shipments and may establish rules for shipping (eg, not shipping perishable goods over a long holiday weekend). Similarly, communication between two electronic devices uses rules, or protocols, that establish how corresponding layers in the two devices interact. In electronic communication, the actual movement of data between two devices occurs only at the physical layer. However, because of the way layered communication works, each layer can be viewed as communicating with the corresponding layer in the other device by means of the established protocols. A layer adds some information to what is going to be communicated and, using the services of the layer, sends it down to the layer below where a similar process may take place. On the receiving side, the information works its way up through the layers with each layer removing the information it needs to perform its functions (Fig 2). In our hypothetical company, the quality control department might add stamps or tags to the product if it meets the quality standards and passes it on to shipping, where it is packaged and placed on a truck. In the second company, the product is unpacked and sent to the quality control (or incoming inspection) department, where the tags or stamps are checked and the product may undergo testing to ensure it has not been damaged in transit. If the product passes inspection, it is sent on for further processing, and final acceptance triggers the accounts payable operation. The protocols used in this process ensure that the shipper has placed appropriate quality control tags on the product so that the receiver can check them against its acceptance documents. If protocol is violated (eg, the shipper does not inspect the product or leaves off the tags), receipt of the product will fail (ie, the receiver will refuse the shipment or may hold it pending resolution of the problem).
One advantage of layered systems in electronic communication is that a layer may be replaced by a newer layer without affecting the other layers. For example, if an alternate physical medium with a transmission speed 10 times greater than that of the existing physical medium becomes available and exactly matches the services and protocols of the existing layer, it may be substituted for the existing layer. If our hypothetical company found a common carrier that delivered freight 50% faster to the same locations, used the same size freight containers, and provided the same insurance coverage as the existing carrier, it could switch carriers without having to modify other operations.
These collections of layers are often called stacks or protocol stacks. The DICOM standard makes use of a layered communications structure and, in fact, is designed to use existing standard communications stacks rather than define its own. One may ask, If DICOM uses existing standards for communication, what then does it do? Why not just use the communications standards as they exist?
2.Tanenbaum AS. Computer networks. 2nd ed. Englewood Cliffs, NJ: Prentice-Hall, 1988.
3.Helsall F. Data communications, computer networks, and open systems. 3rd ed. Wokingham, England: Addison-Wesley, 1992.
<<Previous 1 2 3 4 5 6 7 8 9 Next >>
