ProvBench Traces

Below are provenance traces collected from three applications (Apache, BLAST, PostMark), each run on three operating systems (Linux, Mac OS X, and Windows). Each trace was collected using SPADE with a Linux / Mac OS X / Windows reporter and H2 SQL storage.

For more information about the workloads, please see:

• Hasnain Lakhani, Rashid Tahir, Azeem Aqil, Fareed Zaffar, Dawood Tariq, and Ashish Gehani, Optimized Rollback and Re-computation, 46th IEEE Hawaii International Conference on Systems Science (HICSS), IEEE Computer Society, 2013. [PDF]

The provenance traces are provided in compressed SQL script format. This can be imported into most SQL databases.

	Linux (Linux Audit)	Mac OS X (OpenBSM)	Windows (Process Monitor)
Apache	SQL script, H2 database	SQL script, H2 database	SQL script, H2 database
BLAST	SQL script, H2 database	SQL script, H2 database	SQL script, H2 database
PostMark	SQL script, H2 database	SQL script, H2 database	SQL script, H2 database

Importing into MySQL

To import the provenance traces into MySQL, use the following command on Linux or Mac OS X, replacing <user> and <password> with your MySQL credentials:

gzip -dc < script.tar.gz | mysql -u <user> -p <password>

Viewing the H2 SQL database

To browse the H2 SQL database without importing the SQL script into an existing database, follow the instructions [here](Storing\ provenance).

Generating the provenance traces

For Apache, the following commands were used on Linux and Mac OS X:

./configure
make

The following command was used on Windows:

nmake /f Makefile.win _apached

---

For Postmark, the following configuration settings were used for the benchmark:

Base number of files: 500
Transactions: 500
Files range between 500 bytes and 9.77 kilobytes in size
Block sizes: read=512 bytes, write=512 bytes

---

For BLAST, the input data set was downloaded from ftp://ftp.ncbi.nlm.nih.gov/genomes/INFLUENZA/influenza.faa and the following command was used:

makeblastdb -in influenza.faa -parse_seqids -hash_index -out outputdb