Value-Added Services and Content Platforms (VAS and Content platforms) was carried out with in a group with same name at Tele2 AB in Kista, Stockholm. This group is responsible for network design, capacity planning, dimensioning, Acceptance testing (ATP test), and introducing of new functionality in Tele2's VASplatforms. Acceptance testing is performed on new devices (servers and other network components) in order to verify their capacity and performance guaranteed by their manufactures. Every platform has a guaranteed upper bound performance (based on the license a buyer has paid for), measured by different approaches. For instance for Short Message Service Center (SMSC) platforms, the measurement is based on the maximum number of SMS messages processed per second (SMS/sec), for Multimedia Messaging Service Center (MMSC) platforms the metric is the maximum number of MMS messages processed per second (MMS/sec), and for WAP Gateways it is the maximum number of WAP Transactions Per Second (TPS). This M.Sc. thesis project involved creating two graphical load generators for load testing of SMSC and MMSC platforms. These application-programs are not allowed to occupy unnecessary resources, or cause additional traffic on the radio network (when they are deployed), but they must be powerful enough in order to send and receive traffic in order to derive statistical data about the system's performance. This data will be used for behavioral analysis of these systems, and finally for verifying the guaranteed capacities. These tests are very important and decisive for service providers, who want to be able to offer good quality of service, guarantee availability, and offer reliability. In order to measure the performance and verify the guaranteed performance, two main scenarios were of great importance: Sending 5 messages per second during a interval of 5 minutes. This case will simulate a TV-contest in which the TV audiences submit messages to a predefined number in order to join the contest. Sending 3 Multimedia Messages per second during 30 minutes (for the MMSC performance measurement), and 7 SMS-messages per second during 120 minutes (for the SMSC performance measurement). This case attempts to simulate the traffic that will be generated in the minutes before and after Christmas or New Year. For behavioral analysis and performance measurement of the MMSC and the SMSC platforms an Open Queueing Network model is employed. In this model each server system is considered as a network, consisting of nodes, where each node represents one component inside the system. By considering each node as a single-server queueing system we can take advantage of queueing theory in order to drive several performance results.
Identifer | oai:union.ndltd.org:UPSALLA1/oai:DiVA.org:kth-93112 |
Date | January 2003 |
Creators | Mahdavi, Adrian |
Publisher | KTH, Mikroelektronik och Informationsteknik, IMIT |
Source Sets | DiVA Archive at Upsalla University |
Language | English |
Detected Language | English |
Type | Student thesis, info:eu-repo/semantics/bachelorThesis, text |
Format | application/pdf |
Rights | info:eu-repo/semantics/openAccess |
Page generated in 0.0019 seconds