Face-off: Users do need application-aware networks
Two industry experts debate the merits of application awareness vs. application transparency.
By Issy Ben-Shaul
,
Network World
, 04/17/2006
- Share/Email
- Tweet This
- Print
Organizations today employ widely distributed IP networks that deliver Web-based and client/server applications worldwide.
Even with ample bandwidth, application performance can slow to a snail's pace in this environment, seriously hindering a company's
ability to do business.
To address this challenge, networks need to become more application-aware. That is, intelligent networks must be able to identify
and understand Layer 7 application protocols to optimize or carry out application-specific messages. By identifying types
of application messages, such as high-priority and latency-sensitive signaling messages, or by recognizing message content,
such as a purchase order or stock trade, a network can apply message-specific optimization, security or routing policies.
With greater application-aware functionality, remote offices can experience LAN-like performance, boosting employee productivity.
Application-aware networks also enable IT managers to monitor application performance constantly and optimize bandwidth and
spot trouble early. This enhances application operation and the overall IT infrastructure - all without the huge expense of
building out a network or reworking applications.
For example, a major insurance company began to deploy a homegrown, Web-based application that had taken more than two years
to develop. The application tested well in the data center, so the company was optimistic about delivering an important claims-processing function to its many branches across North
America. The deployment hit a serious snag, however, when employees tried to access the application over their WAN. Application performance slowed to a crawl, and frustrated workers began to miss their quotas. An application that was supposed
to boost productivity reduced it instead, forcing a premature halt to the rollout.
Upgrading the network itself was irrelevant, as bandwidth utilization was less than 20%, and upgrading would have done nothing
to mitigate the latency. Alternatively, the company could have rewritten the claims application to minimize the number of
traversals, but it rejected this option. Instead, it chose to deploy affordable application-aware network devices that understood
its Web protocol. By so doing, the network became more efficient in delivering the performance employees needed. In fact,
application response time improved 2.5 times without changes to the application or user systems.
Comment