Dynamically Scalable Architectures for E-Commerce

1 Motivation E-commerce became an important economic factor during the last years. Approximately 38.5 million customers in Germany have already bought goods or services online. Furthermore the total revenue in Germany is expected to rise from 46 billion Euro in 2006 to 145 billion in 2010. Increasing mobile web usage and the trading firms' expanding international activities boost this development even further (Stahl et al. 2009). This growing acceptance of e-commerce will result in an increasing number of both transactions and customers. Considering the extensive and increasing usage of search agents, web-crawlers and interactive elements in addition, workloads in e-commerce will rise significantly. Moreover, the advancing integration with other systems, like other shops requesting catalogue data, causes additional requests to handle. At the same time, intense competition among retailers tends to result in more demanding requirements with respect to functionality, usability, availability and performance. Rising workloads and more demanding requirements are nowadays handled with statically scaled systems, where the capacity is determined by the expected maximum workload. This results in low resource utilization in data centers conflicting with rising energy costs and CO2 emissions. Fluctuation in workloads impairs the situation even further. After Microsoft released Windows 7, various news channels and blogs reported that many online shops had serious performance problems and some servers broke down completely (Computerbase, 2009; Chronicle, 2009). This highlights the problem that workload forecasts are difficult and that surges of visitors can exceed the expected maximum workload significantly. An examination of a B2B shop system showed