Design of a workload

[jboettcher]

=== I think this might be out of the scope of the seminar, but if we want we could investigate it in future work (guided research/master thesis) ===

In future, we want to evaluate the "changability" of the different layouts: Therefore it would make sense to have a benchmark for it.

You could check if there is already a suitable benchmark in the literature, or define one yourself. It would be nice to have some parameters one could adjust for benchmarking like:

• ratio of edges added
• ratio of nodes added
• ratio of deletions
• ratio of algorithms/queries

[mehemmedv]

Yes, we can better discuss it in your office, when I am back.