Analyzing networks to find bottlenecks has always been a guessing game. Even monitoring tools, which are expensive and reactive, have not necessarily made it easy or possible to find the smoking gun, according to Adam Fogelman, assistant director of
Analyzing networks to find bottlenecks has always been a guessing game. Even monitoring tools, which are expensive and reactive, have not necessarily made it easy or possible to find the smoking gun, according to Adam Fogelman, assistant director of network services at the Cleveland Clinic Foundation, speaking Sunday at the HIMSS meeting.
The solution to bottleneck problems has been to build bigger network pipes, but this may not be the best policy, he said. A blame game often ensues, as network and applications teams point accusing fingers at each other when bottlenecks inevitably occur.
New business drivers, including electronic medical records and the digitization of radiographic images, make it critical that healthcare information is not only available but available immediately and from any location in the enterprise.
"To achieve this goal, hospitals need to control their wide area network bandwidth at the application level," Fogelman said.
Radiology's new business model requires the transfer of large electronic images to and from remote locations, for example. Without proper bandwidth management, critical business applications can be negatively affected in other WAN locations.
Fogelman recommended installing rate, or traffic, shaping policies at all the remote locations, a one-time investment that will yield a positive return by avoiding continual bandwidth expansion. Rate shaping policies (using solutions such as PacketShaper) ensure the performance of mission critical applications by allocating bandwidth based on priorities.
"We use traffic shaping to track application performance and ensure we meet user expectations," Fogelman said.
It's also valuable to help perform capacity planning and modeling when adding new applications, he said.
Traffic shaping has allowed the Cleveland Clinic to:
? eliminate the blame game by permitting analysts to converse with facts, not innuendo
? characterize application behavior of the WAN
? determine WAN requirements for new applications
? proactively plan for needed capacity
? establish and measure service level agreements
"We now have better vision into the network and more efficient use of existing bandwidth," Fogelman said.
The Cleveland Clinic now has the ability to monitor the network continuously at the application level, allowing it to enforce business priorities and forecast capacity growth.