Skip to content

mhausenblas/nosql4lod

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 

Repository files navigation

BNSOL: Benchmarking NoSQL solutions for Linked Data processing

I recently asked: can NoSQL help us in processing Linked Data? Now, let's see where we are.

This is benchmark, BNSOL, is to my knowledge the first attempt to benchmark NoSQL systems concerning Linked Data processing capabilities. It won't be perfect and it won't be comprehensive in the first iteration, but, hey, it's a start ;)

If you want to help out or have an idea what to improve, let me know ...

Candidates

For now, from each NoSQL family one representative is chosen. As an introduction, you might want to check out Rick Cattell's brilliant page on scalable datastores - I'm following his categorisation to name the families:

Why did I choose these? I guess a mixture of how mature and widely used they are as well as some earlier experiences. I plan to add more candidates over time.

Data

I don't know yet, but for the start I'll put together some RDF documents myself and then, once the BNSOL infrastructure (environment + client) is stable, I guess I'll move to something more heavy-weight like the BTC-2010.

Environment

Hm. Either I'm going to use our in-house cluster at DERI or EC2 will have to do the job. Let's see. For now, all I need is my laptop.

Yahoo's YCSB looks like a promising start as well. Alex Popescu has gathered pointers to some platform-specific NoSQL benchmarks as well.

It might make sense to go for a trace-based methodology by leveraging the LOD log data we've gathered for our WebSci10 paper Learning from Linked Open Data Usage: Patterns & Metrics.

License

The software and other artefacts (such as configurations, etc.) provided in the repository are in the Public Domain.

About

BNSOL: Benchmarking NoSQL solutions for Linked Data processing

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published