Microsoft CIFS and MAPI-based applications are lousy network citizens, so if you are about to move your Microsoft servers away from users, you should take pre-emptive measures—or buy extra life insurance to ensure your family will be provided for in case of a user mutiny. In a previous post we described the problem. Now it’s time to describe the fix.
To fix CIFS and MAPI misbehavior you first need a proper application delivery system (ADS) architecture. The architecture should include a portfolio of techniques that fix CIFS and MAPI performance without touching the Microsoft applications themselves. These techniques improve aspects of the applications’ behavior to address the pesky problem of long user task response times. The techniques apply “performance levers” shown in the following formula to transform how CIFS and MAPI-based applications operate over the WAN. To work their magic they:
• Reduce turn counts by streamlining application protocols
• Reduce application payload through content caching and compression
• Improve network protocol behavior and thus increasing available bandwidth
The Application Performance Levers
By properly combining these techniques, you can reduce both payload and turn count by 80 to 90 percent, and dramatically improve performance. The techniques that fix CIFS and MAPI performance are embedded in distributed application delivery systems, also referred to as “symmetrical” WAN application acceleration products. These systems require two devices—one at the data center and the other at the remote site—that work in tandem to transparently improve the behavior of CIFS and MAPI traffic (as well as your other applications).
But acceleration is not enough. You need to add complementary control and reporting to assure consistent, CIFS and MAPI performance over the long haul. Control ensures that CIFS and MAPI performance remains acceptable when there is network congestion, and reporting enables you to monitor, troubleshoot, and further improve CIFS and MAPI-related performance.
The control capability enables you to identify important traffic, tag it for proper downstream handling, and assign it priority treatment through quality-of-service (QoS). The most performance-constraining factor after latency is access bandwidth. A variety of forces from bandwidth-hogging applications and end-of-day traffic surges to DDOS attacks constantly threaten access bandwidth. Without the ability to measure and report on task-level CIFS and MAPI application performance, you are essentially operating your IT ship without a navigation system, and you can’t understand and assess the seriousness of performance problems, much less plot a course to a better future.
You can learn more about these techniques by reading our “Field Guide to Application Delivery Systems” available for free as Report 5085 on our website.
