Benchmarking for WAN-vicious apps, Part 1

Insightful analysis by consultants Steve Taylor and Jim Metzler, plus links to the latest WAN news headlines

Last September we looked at WAN-vicious applications and discussed a couple of specific applications that performed very badly over the WAN. We concluded that it is important for IT organizations to benchmark WAN and application performance. Today, we’ll discuss WAN-vicious applications in greater detail and in the next newsletter we’ll comment on how IT organizations approach benchmarking.

As a point of reference, we consider an application to be WAN-vicious if it runs well over the LAN but poorly over the WAN. One of the reasons why there are so many WAN-vicious applications in production is that in most cases the focus of application development is on delivering the promised software functionality on time, on budget and with relatively few bugs. In only a minority of instances do we find that there is any significant emphasis during the design and development of an application on how well it will run over a WAN. This is true whether the application is developed internally by the IT organization, or acquired from a software vendor.

Transaction based applications tend to have more performance problems than do bulk file transfer apps. In general, the typical transaction requires multiple application turns. There are two sources of the delay that are associated with each application turn. One source is the round-trip latency of the WAN and refers to the amount of time that it takes a bit to travel from the origin to the destination and back again. Unless the WAN is designed in a really convoluted manner, there is not much that can be done to affect this delay as it is dictated by the speed of light. The second is the transmission delay associated with sending whatever information is required. IT organizations can reduce this delay by adding WAN bandwidth.

To continue reading, register here and become an Insider. You'll get free access to premium content from CIO, Computerworld, CSO, InfoWorld, and Network World. See more Insider content or sign in.

Last September we looked at WAN-vicious applications and discussed a couple of specific applications that performed very badly over the WAN. We concluded that it is important for IT organizations to benchmark WAN and application performance. Today, we’ll discuss WAN-vicious applications in greater detail and in the next newsletter we’ll comment on how IT organizations approach benchmarking.

As a point of reference, we consider an application to be WAN-vicious if it runs well over the LAN but poorly over the WAN. One of the reasons why there are so many WAN-vicious applications in production is that in most cases the focus of application development is on delivering the promised software functionality on time, on budget and with relatively few bugs. In only a minority of instances do we find that there is any significant emphasis during the design and development of an application on how well it will run over a WAN. This is true whether the application is developed internally by the IT organization, or acquired from a software vendor.

Transaction based applications tend to have more performance problems than do bulk file transfer apps. In general, the typical transaction requires multiple application turns. There are two sources of the delay that are associated with each application turn. One source is the round-trip latency of the WAN and refers to the amount of time that it takes a bit to travel from the origin to the destination and back again. Unless the WAN is designed in a really convoluted manner, there is not much that can be done to affect this delay as it is dictated by the speed of light. The second is the transmission delay associated with sending whatever information is required. IT organizations can reduce this delay by adding WAN bandwidth.

Some applications require only a handful of application turns for a given transaction but some require a lot more. To exemplify this, assume that an application requires one hundred application turns to complete a particular transaction. If that application is run over a LAN that has 10 microseconds of round-trip delay, then the total round-trip delay for that transaction is 1 millisecond – which is clearly not noticeable. However, if the same application is now run over a WAN that has 120 milliseconds of round-trip delay, the total round-trip delay is 12 seconds – which clearly is very noticeable and qualifies this application as being WAN-vicious.

Read more about lans & wans in Network World's LANs & WANs section.

Steve Taylor is president of Distributed Networking Associates and publisher/editor-in-chief of Webtorials. Jim Metzler is vice president of Ashton, Metzler & Associates.