Hello! I am Neeraja J. Yadwadkar.
This website hosts the details of the work "Discovery of Application Workloads from Network File Traces"(Click here to download the paper) carried by me during my MSc(Engg) at Indian Institute of Science. "Our Group" provides the information about the whole group behind this work.
We aim at designing a trace analysis methodology with Powerful Discriminatory Capabilities as well as Annotation Capabilities. We tackle a specific instance of the problem - given the headers of an Network File System (NFS) trace, to identify the application-level workload that generated it.
We show encouraging results in details using extensive experimentation, the key contributions include:
- Workload Identification
- Trace Annotation
- Trace Sampling
- Automated Learning
- Power of Opcode Sequences
We are here sharing the data that we worked on in this work, which includes basically the NFS traces gathered using a client server configuration located on a 1 Gbps LAN. For more details, please see Resourses->Dataset section. We provide a few python scripts that were used during this work on the NFS Opcode traces, at Resources->Source Code. These are customized scripts, need to be tailored accordingly for their usage further.