The performance of popular Internet Web services is governed by a complex combination of server behavior, network characteristics and client workload - all interacting through the actions of the underlying Transport Control Protocol (TCP). Consequently, even small changes to TCP or to the network infrastructure can have significant impact on end-to-end performance, yet at the same time it is challenging for service administrators to predict what that impact will be. This paper describes the implementation of a tool called Monkey that is designed to help address such questions. Monkey collects live TCP trace data near a server, distills key aspects of each connection (e.g., network delay, bottleneck bandwidth, server delays, etc.) and then is able to faithfully replay the client workload in a new setting.
Did you find this white paper useful?
27 out of 50 users found this white paper useful
Related white papers
Security in Wireless Sensor Networks
Wireless sensor networks are in use for a wide variety of applications: ocean and wildlife monitoring, earthquake monitoring, manufacturing, building safety monitoring, and many military applications. An even wider spectrum...
Layer 3 Malformed Packet Testing in MCI's Public IP Network
Malformed packets associated with routing protocols can cause undesirable consequences if not detected and contained properly as close to the originator as possible. This issue magnifies in scope and complexity...
On Demand Event: Virtualize Application Server Resources to Handle Spikes in Workload Demands
View this Webcast to hear about marketplace trends and get insights into IBM customer deployment of virtualization technology and how it can apply to the application layer of your architecture.
Best Practices - Monitoring and Controlling Peer-to-Peer (P2P) Applications
Peer to Peer file sharing steals bandwidth that should otherwise be used for business-critical applications. When bandwidth is consumed by P2P applications, the business operations that are tied to those...
Detecting and Resolving Locking Problems Using BMC MAINVIEW for DB2
The presenter of this webcast will discuss the fundamentals of DB2 locking and show how BMC MAINVIEW for DB2 can help identify and resolve DB2 locking problems.
The Real Cost of Spam
Unsolicited Commercial E-mail, or spam, is not just an annoying problem. It's a costly one. This paper examines the different cost impacts, both direct and indirect, that spam has on...
Top Ten Misconceptions about Performance & Availability Monitoring
Learn about 10 aspects of monitoring that may have once been perfectly sound best practices, but now no longer apply in most availability and performance monitoring situations.
