Back to basics: Duplex and symmetry


Photo courtesy of DeClanTM

Over the past few weeks I've been onsite helping some customers update their network architectures and network management strategies. During these discussions I was reminded that since many network administrators get their technical educations through a combination of on-the-job training (OJT), classroom training, and self-study there can sometimes be a few gaps in their technical knowledge. This is especially true for areas of expertise that aren't commonly discussed or are thought of as being so fundamental that they don't warrant a discussion. This week we're going to cover a few different concepts that are commonly missed but very important to understand.


This is commonly used as a slang term for a network connection, especially one that is used to connect disparate networks or to connect your network to the internet as a "pipe". This slang is derived from an analogy that many of us use when explaining network concepts where we compare network connections to water pipes. In this analogy, bandwidth would be the diameter of the pipe and latency would be the amount of time that it takes water to travel the length of the pipe. What's interesting is that while we all label a connection like this as a "pipe", in actuality each connection is a set of "pipes" - one going in each direction, to and from the network.

The term duplex is used to describe whether the link can both send and receive data at the same time (full duplex) or whether it has to alternate sending and receiving (half duplex). Nearly all modern network connections are full duplex and so this concept is rarely discussed today. However, understanding that there are actually separate connections for inbound and outbound traffic, with their own capacities for bandwidth and their own factors affecting latency, is an important part of understanding how data networks work.


Now that we've discussed the fact that most connections are full duplex (pipes for each direction), let's talk about symmetry. A symmetrical connection has the same sized pipe or the same available bandwidth in each direction. While it's commonly assumed that most connections are symmetrical, this is rarely true of fee -based connections. In other words, LANs are usually symmetrical; however, WANs seldom are. A very common example of this would be your broadband connection at home. For instance, my connection is 24 Mbps for inbound traffic and 6 Mbps for outbound traffic. These asymmetrical connections represent the connections that we're most commonly called in to troubleshoot, that we're constantly trying to shape and optimize, and that we're usually bickering over SLAs on.

Understanding these concepts and their effects on network traffic and application performance is an important part of your network management strategy. Next time we'll discuss network adjacencies and the differences between ingress and egress traffic types.

Have a common technology term that you think is commonly misunderstood? Do you find yourself explaining some of the same concepts over and over? Leave a comment and share your thoughts...


Copyright © 2012 IDG Communications, Inc.

It’s time to break the ChatGPT habit
Shop Tech Products at Amazon