database - how to understand experiment performance for partition databse -
i studying paper fast distributed transactions partitioned database systems , found paragraph did not understand. difference between node /machine /partition
, meaning of dist warehouse?
figure 4 shows total , per-machine throughput (tpc-c new order transactions executed per second) function of number of calvin nodes, each of stores database partition containing 10 tpc-c warehouses. investigate calvin’s handling of distributed transactions, multi-warehouse new order transactions (about 10% of total new order transactions) access second warehouse not on same machine first. because each partition contains 10 warehouses , new order updates 1 of 10 “districts” warehouse, @ 100 new order transactions can executing concurrently @ machine (since there no more 100 unique districts per partition,
Comments
Post a Comment