"System Design Issues for Internet Middleware Services: Deductions from a Large Client Trace"

	This paper presents a detailed analysis of a large client-side web trace. The paper 
demonstrates the heterogeneity of user population, the existence of a strong and very predictable 
diurnal cycle in the users' web activity, the burstiness of user requests at fine-grained time 
scales, the presence of locality in the user requests, and the high latency that services encounter 
when delivering data to clients, implying that services will need to maintain a very large number 
of simultaneously active requests.

	TranSend middleware service addresses several of the issues discovered. TranSend service runs
on a cluster of machines. User requests are load balanced across the cluster in order to maximize
request throughput and minimize the end-to-end latency of each request to address burstiness issue.

	This paper was published in 1997. Heterogeneity of user population, existence of predictable cycles
in activity, burstiness of user requests are not new information anymore. Much research has been done
in these areas since 1997. This paper is supported by a concrete analysis and strong numbers.