What is the relationship between application errors and network performance?

What is the relationship between application errors and network performance?

What is the relationship between application errors and network performance?

    Requires Free Membership to View

    Login

    By submitting your registration information to SearchNetworking.com you agree to receive email communications from TechTarget and TechTarget partners. We encourage you to read our Privacy Policy which contains important disclosures about how we collect and use your registration and other information. If you reside outside of the United States, by submitting this registration information you consent to having your personal data transferred to and processed in the United States. Your use of SearchNetworking.com is governed by our Terms of Use. You may contact us at webmaster@TechTarget.com.

Packet loss and end-to-end latency are the major issues that affect application performance in today's enterprise networks. CRC errors are no longer a significant source of packet loss as most Ethernet connections are switched versus shared (i.e. CSMA) so most packet loss is either related to network level buffering issues due to oversubscription or link quality. In either case, the packet loss directly affects TCP performance which in turn directly affects application performance. Additionally, high latency between the application and the user, such as what might be found on a low bandwidth WAN link, directly affects TCP performance due to its affect on the TCP windowing algorithm. In many cases, a WAN link can suffer both from high latency and packet loss, rendering these links unusable for many enterprise class applications. There are solutions available such as an Application Delivery Gateway and WAN optimizers that can help optimize TCP performance to counter the affects of packet loss and increase application availability.

Optimizing TCP performance in the network is, however, only one part of the equation. There are no set rules for application performance and typically application performance requirements will be highly dependent on the types of application being delivered and the needs of the organization. For example, a company with a very mobile employee base may have remote access needs requiring that the application delivery infrastructure not only effectively deliver an application to users on a LAN connection, but over many different types of WAN connections. The key is to provide the same user experience as though the remote employee was connected to the LAN.

When determining application performance needs and setting criteria for an application or group of applications, the network and/or application designer must understand the network, application goals, user base and the requirements of the organization. Additionally, it's important to have good monitoring tools that can help distinguish between application performance and WAN performance. This is a difficult thing to achieve and often requires having some knowledge of the details of a particular application, so a turnkey solution probably doesn't exist.

This was first published in January 2006

Back to top