Brandon Holt edited Evaluation.md  over 9 years ago

Commit id: 9b409000f10456c82286ebabdd643d4d77e6d896

deletions | additions      

       

- *Goals*  - show that this approach does better than OCC/contending  - show that it scales out in terms of increasing contention and number of nodes  - show that  the technique subsumes existing work and opens the door for avoiding contention  -  new kinds of efficient transactions workloads enabled by this? (may just be made significantly more efficient)  - graph database  - must we support multi-node transactions? is it worth it?  - transactions: