Hierarchical Characterization and Generation of Blogosphere Workloads

We present a thorough characterization of the access patterns in blogspace, which comprises a rich interconnected web of blog postings and comments by an increas- ingly prominent user community that collectively dene what has become known as the blogosphere. Our characterization of over 35 million read, write, and manage- ment requests spanning a 28-day period is done at three dieren t levels. The user view characterizes how individual users interact with blogosphere objects (blogs); the object view characterizes how individual blogs are accessed; the server view char- acterizes the aggregate access patterns of all users to all blogs. The more-interactive nature of the blogosphere leads to interesting trac and communication patterns, which are dieren t from those observed for traditional web content. We identify and characterize novel features of the blogosphere workload, and we show the similarities and dierences between typical web server workloads and blogosphere server work- loads. Finally, based on our main characterization results, we build a new synthetic blogosphere workload generator called GBLOT, which aims at mimicking closely a stream of requests originating from a population of blog users. Given the increas- ing share of blogspace trac, realistic workload models and tools are important for capacity planning and trac engineering purposes.